Promotional position matching method and device, promotional information display method and device

Abstract

The present application relates to a promotion position matching method and device, and a promotion information display method and device. The promotion position matching method includes: determining the promotion position and candidate promotion information corresponding to the user identifier; obtaining the estimated exposure click rate of the candidate promotion information; generating the estimated exposure click rate according to the user characteristics corresponding to the user identifier; determining the bidding score corresponding to the candidate promotion information according to the estimated exposure click rate and the unit click value of the candidate promotion information; screening the promotion information whose corresponding bidding score meets the promotion conditions from the candidate promotion information; matching the promotion position with the promotion information source to which the screened promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source. The solution provided by the present application enables the promotion position resources to be fully utilized and avoid waste.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a promotion position matching method and device, and a promotion information display method and device. Background Art

[0002] With the development of computer technology, more and more promotional information is presented to users online. For example, in the field of online advertising display, advertising trading platforms usually allocate advertising traffic resources of SSP (Supply-Side Platform) based on the characteristics and demand of the advertising information to be delivered in DSP (Demand-Side Platform), so that the advertising traffic resources allocated to DSP can bring advertising value to DSP and bring corresponding advertising traffic resource value to SSP.

[0003] However, currently, for promotional positions to be placed for promotional information, most of them are allocated to various promotional information sources in a fixed ratio, resulting in a waste of promotional position resources. Summary of the invention

[0004] Based on this, it is necessary to provide a promotion position matching method, apparatus, computer-readable storage medium and computer device to address the technical problem that the existing way of allocating promotion positions leads to waste of promotion position resources.

[0005] A promotion bit matching method, the method comprising:

[0006] Determine the promotion location and candidate promotion information corresponding to the user identifier;

[0007] Obtaining an estimated exposure click rate of the candidate promotion information; the estimated exposure click rate is generated according to a user feature corresponding to the user identifier;

[0008] Determine a bidding score corresponding to the candidate promotion information according to the estimated exposure click rate and the unit click value of the candidate promotion information;

[0009] Filtering promotion information whose corresponding bidding scores meet promotion conditions from the candidate promotion information;

[0010] The promotion position is matched with the promotion information source to which the filtered promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source.

[0011] A promotion information display method, the method comprising:

[0012] Determine the promotion location corresponding to the user ID;

[0013] Receive promotion information matching the promotion position; the promotion information comes from a promotion information source matching the promotion position; the promotion information source is a promotion information source to which candidate promotion information corresponding to the user identifier and whose corresponding bidding score meets the promotion conditions belongs; the bidding score is determined based on an estimated exposure click rate of the candidate promotion information and a unit click value of the candidate promotion information; the estimated exposure click rate is generated based on user features corresponding to the user identifier;

[0014] The matched promotion information is displayed in the promotion position.

[0015] A promotion position matching device, the device comprising:

[0016] A determination module, used to determine a promotion location and candidate promotion information corresponding to a user identifier;

[0017] An acquisition module, used to acquire an estimated exposure click rate of the candidate promotion information; the estimated exposure click rate is generated according to a user feature corresponding to the user identifier;

[0018] A bidding score determination module, used to determine the bidding score corresponding to the candidate promotion information according to the estimated exposure click rate and the unit click value of the candidate promotion information;

[0019] A screening module, used to screen the promotion information whose corresponding bidding scores meet the promotion conditions from the candidate promotion information;

[0020] The matching module is used to match the promotion position with the promotion information source to which the screened promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source.

[0021] A promotion information display device, comprising:

[0022] A promotion location determination module, used to determine a promotion location corresponding to a user identifier;

[0023] a promotion information receiving module, configured to receive promotion information matching the promotion position; the promotion information comes from a promotion information source matching the promotion position; the promotion information source is a promotion information source to which candidate promotion information corresponding to the user identifier and whose corresponding bidding score satisfies the promotion condition belongs; the bidding score is determined based on an estimated exposure click rate of the candidate promotion information and a unit click value of the candidate promotion information; the estimated exposure click rate is generated based on user features corresponding to the user identifier;

[0024] A display module is used to display the matched promotion information in the promotion position.

[0025] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the processor executes the steps of the above-mentioned promotion position matching method or promotion position display method.

[0026] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the above-mentioned promotion position matching method or promotion position display method.

[0027] The above-mentioned promotion position matching method, device, computer-readable storage medium and computer equipment, promotion information display method, device, computer-readable storage medium and computer equipment match the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click-through rate of the candidate promotion information, and the estimated exposure click-through rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click-through rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 A diagram of an application environment for promoting a bit matching method in an embodiment;

[0029] Figure 2 A schematic diagram of a flow chart of a method for generalizing bit matching in one embodiment;

[0030] Figure 3 A framework diagram for matching promotional positions with promotional information sources in one embodiment;

[0031] Figure 4 An example diagram of a discovery home page tab of an App installed on a terminal for downloading application programs in one embodiment;

[0032] Figure 5 A schematic diagram of a method for calculating the unit click value of the third type of promotional information in one embodiment;

[0033] Figure 6 It is a schematic diagram of a process of calculating the initial exposure weight corresponding to the promotion information source in one embodiment;

[0034] Figure 7 A schematic diagram of a flow chart of a method for generalizing bit matching in a specific embodiment;

[0035] Figure 8 A schematic diagram of a flow chart of a promotion information display method in one embodiment;

[0036] Fig. 9 A structural block diagram of a generalized bit matching device in one embodiment;

[0037] Fig.10 A structural block diagram of a generalized bit matching device in another embodiment;

[0038] Fig.11 It is a structural block diagram of a generalized bit matching device in another embodiment;

[0039] Fig.12 It is a structural block diagram of a promotion information display device in one embodiment;

[0040] Fig.13 is a structural block diagram of a computer device in one embodiment;

[0041] Fig.14 FIG. 4 is a structural block diagram of a computer device in one embodiment. DETAILED DESCRIPTION

[0042] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0043] Figure 1 FIG. 1 is a diagram of an application environment for promoting a bit matching method in an embodiment. Figure 1 The promotion position matching method is applied to a promotion position matching system. The promotion position matching system includes a terminal 110 and a server 120. The terminal 110 and the server 120 are connected via a network. The terminal 110 may be a fixed terminal or a mobile terminal, and the mobile terminal may be at least one of a mobile phone, a tablet computer, a laptop computer, etc. The server 120 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0044] like Figure 2 As shown, in one embodiment, a generalized bit matching method is provided. This embodiment mainly applies the method to the above Figure 1 The server 120 in FIG. Figure 2 The promotion position matching method specifically includes the following steps:

[0045] S202: Determine a promotion position and candidate promotion information corresponding to the user identifier.

[0046] Among them, promotion information is the online communication medium of the promotion object. The promotion information can be text, pictures, keywords, rankings or videos. The computer device can use text description, picture display or video presentation to show the object to be promoted to the user. The promotion object can be, for example, an application to be downloaded, a product to be purchased, etc.

[0047] Candidate promotion information is a collection of promotion information corresponding to a user identifier. The types and quantities of candidate promotion information corresponding to different user identifiers may be different. In one embodiment, the candidate promotion information belongs to different promotion information sources. The promotion information source is the party that demands promotion positions. There may be multiple types of promotion information sources. For example, the promotion information source includes advertising parties, game parties, basic software parties, etc.

[0048] Promotional slots are slots on the Internet media used to display promotional information. Promotional slots can be advertising slots used to display various applications on Apps (Applications, computer applications) used to download applications. For example, the recommended list of "recommended applications" displayed in the home page tab of an App used to download applications. Promotional slots can also be advertising slots in social applications, scrolling advertising bars on portal websites, contextual recommendation tabs, search result lists in search engines, and so on.

[0049] In one embodiment, after detecting the user's exposure behavior, the terminal may send the user identifier and the promotion position to be allocated corresponding to the user identifier to the server. After receiving the user identifier, the server may obtain the promotion position to be allocated and candidate promotion information corresponding to the user identifier.

[0050] The user's exposure behavior may be, for example, the user logging into an App used to download applications, or opening the App; it may also be the user browsing the social circle of a social application; it may also be the user opening a web page of a portal website, entering search keywords in a search engine, and so on.

[0051] The user identifier is a character used to uniquely identify a user. For example, it can be a user's mobile phone number, a user account, etc. In one embodiment, the user identifier can also be a terminal identifier corresponding to any terminal. For example, for a mobile terminal, the user identifier can be the IMEI (International Mobile Equipment Identity) of the mobile terminal.

[0052] In one embodiment, the server can extract user behavior characteristics from the exposure behavior of the user corresponding to the user identifier on various Internet media, obtain promotion information characteristics of each promotion information, and generate candidate promotion information corresponding to the user identifier based on the matching degree between the extracted user behavior characteristics and the promotion information characteristics of each promotion information.

[0053] S204, obtaining an estimated exposure click rate of the candidate promotion information; the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier.

[0054] Among them, the predicted click-through rate (PCTR) is the probability that the candidate promotion information will be clicked after being presented to the user through the promotion position corresponding to the user identifier.

[0055] Specifically, after obtaining the candidate promotion information corresponding to the user identifier, the server can obtain the estimated exposure click rate of the candidate promotion information according to the user characteristics corresponding to the user identifier. The server can also obtain the estimated exposure click rate of the candidate promotion information according to the promotion information characteristics of each promotion information in the candidate promotion information and the user characteristics.

[0056] User features may include user portrait features, user behavior features, and cross-combination features of user features and promotion information features. Among them, user portrait features include the age, gender, height, account information of the user corresponding to the user ID, the network environment and time environment of the terminal corresponding to the user ID, etc. User behavior features include the historical exposure behavior of the user corresponding to the user ID in the candidate promotion information, such as historical click behavior for the candidate promotion information, historical download behavior corresponding to the object promoted by the candidate promotion information, historical search behavior and historical consumption behavior, etc. The cross-combination feature of user features and promotion information features refers to the combined attributes of various aspects of the mined candidate promotion information and user ID, such as the number of matches between the keywords searched by the user and the words in the content of the candidate promotion information, etc. Since the estimated exposure click rate is generated based on the user features corresponding to the user ID, it can fully reflect the matching situation between the candidate promotion information and the user.

[0057] The promotion information features may include content features of the promotion information, including but not limited to the title text, price, sales volume, attributes of the promotion object, category of the promotion object, and the promotion information source to which the promotion information belongs.

[0058] In one embodiment, when the promotion object corresponding to the candidate promotion information is an application to be downloaded, the estimated exposure click rate of the candidate promotion information obtained by the server can be used to estimate the probability that the user corresponding to the user identifier clicks on the displayed promotion information and downloads the application.

[0059] S206: Determine the bidding score corresponding to the candidate promotion information according to the estimated exposure click rate and the unit click value of the candidate promotion information.

[0060] Among them, the cost per click (CPC) is the value that can be brought by a user clicking once after the candidate promotion information is displayed in the promotion position corresponding to the user ID. The value of the promotion position can be quantified by the product of the number of clicks and the cost per click, that is:

[0061] income = n × cpc, where n represents the number of times the promotional information displayed in the promotional position is clicked, cpc represents the unit click value of the displayed promotional information, and income represents the value of the promotional position.

[0062] In one embodiment, the setting standards of the unit click value corresponding to the candidate promotion information from different promotion information sources may be different. The unit click value corresponding to the candidate promotion information from the same promotion information source may also be different when the promotion objects are different. The unit click value of the candidate promotion information may reflect the value advantage of the promotion information among all the candidate information.

[0063] The bidding score is a reference value for matching candidate promotion information with the promotion position corresponding to the user identifier. The bidding score of the candidate promotion information can be the product of the estimated exposure click rate and the unit click value of the candidate promotion information, or it can be the linear weighted sum of the estimated exposure click rate and the unit click value calculated under their respective weighting coefficients. When matching promotion positions based on the bidding score, the user's demand for promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0064] Specifically, after obtaining the candidate promotion information corresponding to the user identifier, the server may obtain the single click value corresponding to the candidate promotion information, and determine the bidding score of the candidate promotion information based on the estimated exposure click rate and unit click value of the candidate promotion information.

[0065] In one embodiment, after obtaining the candidate promotion information identified by the user, the server may determine the promotion information source to which the candidate promotion information belongs, and determine the unit click value of each candidate promotion information according to the unit click value standards set for different promotion information sources.

[0066] In one embodiment, the step of determining the unit click value of the candidate promotion information includes: when the candidate promotion information belongs to the first type of promotion information, obtaining the bid value of a single click of the candidate promotion information; and determining the bid value as the unit click value of the candidate promotion information.

[0067] The source of the promotion information of the first type of promotion information may be an advertiser. Specifically, when the server determines that the source of the promotion information to which the candidate promotion information belongs is an advertiser, the server obtains the bid value of the advertiser for a single click on the candidate promotion information, and determines the obtained bid value as the unit click value of the candidate promotion information. For example, when the promotion object promoted by the candidate promotion information is an online shopping application to be downloaded, the corresponding unit click value is the bid value paid by the advertiser for downloading the online application once.

[0068] In one embodiment, the step of determining the unit click value of the candidate promotion information includes: when the candidate promotion information belongs to the second type of promotion information, obtaining the average revenue value of the application downloaded after the candidate promotion information is clicked within a specified time period; and determining the unit click value of the candidate promotion information based on the average revenue value.

[0069] Among them, the promotion information source of the second type of promotion information may be the game party. Specifically, when the server determines that the promotion information source to which the candidate promotion information belongs is the game party, and the promotion object promoted by the candidate promotion information is the game application to be downloaded, the average revenue value of the downloaded game application within a specified time period after the candidate promotion information is clicked is obtained, and the obtained average revenue value is determined as the unit click value of the candidate promotion information. For example, the server can estimate the life cycle value (Life-Time-Value 14, LTV14) of the game application within 14 days after downloading based on the user's historical payment record for the game application. The server can also use m*LTV14 (m=1,2,3,…,n) as the unit click value of the candidate promotion information.

[0070] S208, selecting promotion information whose corresponding bidding scores meet the promotion conditions from the candidate promotion information.

[0071] Specifically, after determining the bidding scores of the candidate promotion information corresponding to the user identifier, the server selects the promotion information whose corresponding bidding scores meet the promotion conditions from the candidate promotion information. The promotion conditions met by the selected promotion information may be the highest bidding score, the bidding score that meets the preset conditions (ranked at the top) and the highest estimated exposure rate, or the bidding score that meets the preset conditions and the highest unit click value, etc.

[0072] S210, matching the promotion position with the promotion information source to which the filtered promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source.

[0073] Specifically, after the server filters out the promotion information whose bidding scores meet the promotion conditions from the candidate promotion information, it determines the promotion information source to which the filtered promotion information belongs, and matches the promotion position corresponding to the user identifier with the promotion information source to which the filtered promotion information belongs. The promotion position corresponding to the user identifier can display the promotion information from the matched promotion information source, thereby achieving matching of the promotion position with the promotion information source based on the bidding score.

[0074] In one embodiment, the promotion position matching method also includes the following steps: receiving promotion information from the matched promotion information source; the received promotion information complies with the bidding sorting method of the matched promotion information source; pushing the received promotion information, and the pushed promotion information is used for display in the matched promotion position.

[0075] In this embodiment, the promotional information received and displayed in the matching promotional position is not necessarily the promotional information with the highest bidding score that has been screened out, but the promotional information selected by the promotional information source according to an internal bidding sorting method, thereby realizing the separation of the promotional position matching mechanism and the internal sorting mechanism of the promotional information source.

[0076] In one embodiment, the server may match the promotion location corresponding to the user identifier with the filtered promotion information that meets the promotion condition, so that the promotion location is used to display the filtered promotion information.

[0077] like Figure 3 As shown, it is a framework diagram for matching the promotion position with the promotion information source in one embodiment. Among them, the promotion position is a slot in the recommendation list in the home page label of the App for downloading the application program. It can be understood that the promotion information displayed in the slot of the recommendation list is used to promote the corresponding App. The promotion information source includes three business access parties, namely the advertising business access party, the game business access party and the basic software business access party. The promotion objects promoted by the corresponding promotion information of each business access party are advertising App, game App and basic software App respectively. After the slot allocation service detects that the user opens the home page label of the App for downloading the application program installed on the terminal, for each slot in the recommendation list in the home page label, the slot allocation service obtains the candidate promotion information corresponding to the user identifier, calculates the corresponding bidding score for each candidate promotion information, and selects the business access party to which the promotion information with the highest bidding score belongs from the candidate promotion information, and allocates the slot to the business access party. The business access party can allocate the slot to any promotion information according to its internal sorting mechanism, and the terminal displays the promotion information in the slot for promoting the corresponding App, such as A, B and C.

[0078] like Figure 4FIG. 1 is an example diagram of a discovery home page tab of an App installed on a terminal for downloading applications in one embodiment. Figure 4 The example diagram shows that there are multiple slots in the recommendation list, and each slot is used to display the corresponding promotion information and promote the corresponding App.

[0079] The above-mentioned promotion position matching method matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click rate of the candidate promotion information, and the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0080] In one embodiment, obtaining an estimated exposure click rate of candidate promotion information includes: obtaining user characteristics corresponding to a user identifier; extracting promotion information characteristics of the candidate promotion information; generating a feature vector of the candidate promotion information based on the user characteristics and the promotion information characteristics; and determining the estimated exposure click rate of the candidate promotion information through an exposure click rate prediction model corresponding to a promotion information source to which the candidate promotion information belongs and based on the feature vector of the candidate promotion information.

[0081] For the explanation of the user features corresponding to the user identifier and the promotion information features of the candidate promotion information, please refer to the above description, which will not be repeated here. The feature vector is the relationship between the candidate promotion information and the user corresponding to the user identifier, and the mathematical expression of each feature of the candidate promotion information. The dimension of the feature vector is related to the number of user features and promotion information features referenced.

[0082] Specifically, after obtaining the user features corresponding to the user identifier and the promotion information features of the candidate promotion information, the server maps each feature to a distributed representation space, generates a feature vector corresponding to the candidate promotion information, and inputs the feature vector into a trained exposure click rate prediction model corresponding to the promotion information source to which the candidate promotion information belongs. After calculation, the server outputs the estimated exposure click rate of the candidate promotion information.

[0083] In one embodiment, the exposure click rate prediction model may be a machine learning model, such as a logistic regression (LR) model.

[0084] In this embodiment, the user characteristics and promotion information characteristics corresponding to the user ID are expressed in the form of feature vectors, which can fully reflect the matching situation between the candidate promotion information and the user. The estimated exposure click rate predicted by the trained machine learning model can truly reflect the probability that the candidate promotion information may be clicked by the user with the user ID.

[0085] In one embodiment, the training steps of the exposure click rate prediction model include: obtaining historical user characteristics, historical promotion information characteristics and sample labels corresponding to the historical promotion information belonging to the promotion information source; generating a historical feature vector corresponding to the promotion information belonging to the promotion information source based on the historical user characteristics and the historical promotion information characteristics; inputting the historical feature vector into the exposure click rate prediction model to obtain the historical exposure click rate corresponding to the historical feature vector; when it is determined that the training stop condition is not met based on the historical exposure click rate and the sample label, adjusting the model parameters of the exposure click rate prediction model and continuing training until the training stop condition is met.

[0086] Specifically, for each promotion information source, the server obtains the historical promotion information belonging to the corresponding promotion information source, uses the historical promotion information as sample data, and obtains the exposure click rate prediction model corresponding to each promotion information source through model training. Among them, the historical promotion information includes positive samples and negative samples. Positive samples are samples that have been clicked, and negative samples are samples that have not been clicked. Positive samples and negative samples correspond to different sample labels.

[0087] In one embodiment, after obtaining the historical promotion information of the corresponding promotion information source, the server extracts historical user characteristics and historical promotion information characteristics from the historical promotion information to generate a historical feature vector corresponding to the promotion information belonging to each promotion information source, and inputs the historical feature vector into the exposure click rate prediction model to obtain the historical exposure click rate.

[0088] In one embodiment, the server may compare the obtained historical exposure click rate with a preset threshold, and then compare the comparison result with the sample label of the promotion information to determine whether the historical exposure click rate is consistent with the actual situation. For example, the preset threshold is 0.5, the actual sample label of the positive sample is 1, and the actual sample label of the negative sample is 0. When the historical exposure click rate corresponding to the historical promotion information output by the exposure click rate prediction model is less than 0.5, the sample label predicted for the historical promotion information is 0, and the predicted sample label is compared with the actual sample label. If they are the same, it is determined that the prediction is correct, and if they are different, it is determined that the prediction is wrong; similarly, when the historical exposure click rate corresponding to the historical promotion information output by the exposure click rate prediction model is greater than or equal to 0.5, the sample label predicted for the historical promotion information is 1, and the predicted sample label is compared with the actual sample label. If they are the same, it is determined that the prediction is correct, and if they are different, it is determined that the prediction is wrong.

[0089] In one embodiment, the server may adjust the model parameters by counting the accuracy of click-through rate prediction for historical promotion information in the sample data until the accuracy meets the training stop condition, then stop training to obtain an exposure click-through rate prediction model for each historical promotion information source.

[0090] In the above embodiment, by acquiring historical promotion information of each promotion information source and extracting historical user features and historical promotion information features from real historical data to train the model, the accuracy of the exposure click rate prediction model of each promotion information source can be guaranteed.

[0091] In one embodiment, the step of determining the unit click value of the candidate promotional information includes: when the candidate promotional information belongs to the third type of promotional information, determining the estimated unit click value of the third type of promotional information; the estimated unit click value is determined when the total revenue value of the first promotional page and the total revenue value of the second promotional page are equal; the promotion position of the first promotional page is only used to display the third type of promotional information; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; the estimated unit click value is determined as the unit click value of the candidate promotional information.

[0092] Among them, the promotion information source of the third type of promotion information can be the basic software party. Specifically, when the server determines that the promotion information source to which the candidate promotion information belongs is the basic software party, and the promotion object promoted by the candidate promotion information is the basic application to be downloaded, the candidate promotion information has no unit click value, but there is a situation where the user clicks on the candidate promotion information and downloads the promoted application. When the candidate promotion information is displayed in the promotion position of the promotion page, it will cause the value of the promotion position on the page that displays non-third type promotion information to be reduced. The estimated unit click value of the third type of promotion information from the basic software party can be calculated based on this reduced value.

[0093] Specifically, if the promotional position of the first promotional page is only used to display the third type of promotional information, then the total revenue value of the first promotional page is the sum of the values ​​of all promotional positions displaying the third type of promotional information on the page; if the promotional position of the second promotional page is used to display the promotional position of the third type of promotional information and the promotional position displaying non-third type of promotional information, then the total revenue value of the second promotional page is the sum of the values ​​of all promotional positions displaying the third type of promotional information and the values ​​of all promotional positions displaying non-third type of promotional information on the page. When the total revenue value of the first promotional page is equal to the total revenue value of the second promotional page, the depreciation value of the promotional position displaying non-third type of promotional information on the second promotional page can be calculated, and the estimated unit click value of the third type of promotional information can be obtained based on the depreciation value, and the server can determine the estimated unit click value as the unit click value of the candidate promotional information.

[0094] In this embodiment, by obtaining the value loss of non-third type promotion information displayed in a promotion page with the same total revenue value, the potential value of the third type promotion information after being displayed in the promotion position in the promotion page can be expressed.

[0095] In one embodiment, the step of determining the unit click value of the candidate promotional information includes: when the candidate promotional information belongs to the third type of promotional information, obtaining the first number of clicks of the third type of promotional information in the first promotional page; the promotion position of the first promotional page is only used to display the third type of promotional information; obtaining the second number of clicks of the third type of promotional information and the revenue value of non-third type promotional information in the second promotional page; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; dividing the obtained revenue value by the difference between the first number of clicks and the second number of clicks to obtain the unit click value of the candidate promotional information.

[0096] Specifically, the unit click value of the third type of promotional information can be derived by the following formula:

[0097] Click1×cpc3=Click2×cpc3+Income (Formula 1)

[0098] The left side of formula 1 represents the revenue value of the promotion position in the first promotion page, and the right side represents the revenue value of the promotion position in the second promotion page. Click1 represents the first number of clicks; cpc3 represents the unit click value of the third type of promotion information; Income represents the revenue value of the promotion position that displays non-third type promotion information in the second promotion page; Click2 represents the second number of clicks. The unit click value of the third type of promotion information can be derived from formula 1:

[0099]

[0100] In one embodiment, when the third type of promotion information promotes an application to be downloaded, Click1 in Formula 2 represents the first download count of each application to be downloaded in the first promotion page; Income represents the revenue value of the promotion position displaying non-third type promotion information in the second promotion page; Click2 represents the second click count of each application to be downloaded in the second promotion page. Formula 2 can be used to obtain the single click value corresponding to the third type of promotion information when promoting the application to be downloaded.

[0101] In one embodiment, the unit click value of the third type of promotion information may be the same. The server may pre-store the unit click value of the candidate promotion information belonging to the third type of promotion information, and directly obtain the corresponding unit click value after determining the promotion information source to which the candidate promotion information belongs.

[0102] like Figure 5 FIG. 1 is a schematic diagram of a method for calculating the unit click value of the third type of promotional information in an embodiment. Figure 5 , the promotion information of the two promotion pages promotes the application to be downloaded; in one of the promotion pages, it is a purely distribution-oriented traffic, indicating that the promotion page is only used to display the third type of promotion information, and only brings downloads of the application to be downloaded; in the other promotion page, it is a purely revenue-oriented traffic, indicating that the promotion page can display non-third type promotion information, and at the same time bring downloads and revenue of the application to be downloaded; the traffic value of the two promotion pages is conserved, which means that the revenue values ​​of the promotion positions displaying the promotion information in the two promotion pages are equal; the traffic value corresponding to each of the two promotion pages can be determined by the downloads of the promoted application to be downloaded and the estimated unit click value, so that the unit click value of the third type of promotion information can be derived based on the estimated unit click value.

[0103] In this embodiment, based on the principle that the revenue values ​​of the same promotion page are equal, the method for obtaining the candidate promotion information belonging to the third type of promotion information can be determined after transformation, and the corresponding unit click value can be determined according to the number of clicks on the candidate promotion information.

[0104] In one embodiment, the bidding score is determined based on the estimated exposure click rate, the unit click value, and the exposure weight of the promotion information source to which the candidate promotion information belongs.

[0105] For example, the bidding score of the candidate promotion information can be the product of the estimated exposure click-through rate of the candidate promotion information, the unit click value and the exposure weight of the promotion information source to which the candidate promotion information belongs. It can also be the linear weighted sum of the estimated exposure click-through rate, the unit click value and the exposure weight of the promotion information source to which the candidate promotion information belongs calculated under their respective weighting coefficients.

[0106] In one embodiment, the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs includes: counting the exposure ratio of the promotion information source to which the candidate promotion information belongs after the previous promotion position matching; and determining the exposure weight of the promotion information source to which the candidate promotion information belongs when the promotion position matching is performed this time based on the counted exposure ratio.

[0107] The exposure ratio is the ratio of the promotion information displayed on the promotion page that belongs to the same promotion information source to all promotion information. Specifically, after the previous promotion position matching, the server can count the exposure ratio of each promotion information source after the previous promotion position matching, and use the previous exposure ratio as the exposure weight of the promotion information source to which the candidate promotion information belongs when matching the promotion position this time.

[0108] For example, there are 10 promotion positions corresponding to the user ID in the promotion page. When the promotion positions are matched last time, for these 10 promotion positions, the server will obtain the candidate promotion information corresponding to the user ID, and obtain the bidding score of the candidate promotion information based on the product of the estimated exposure click rate, unit click value and exposure weight of the candidate promotion information, and select the promotion information whose bidding score meets the promotion conditions from the candidate promotion information, and match each promotion position with the promotion information source to which the selected promotion information belongs. For example, 3 of the 10 promotion positions are matched to the promotion information source to which the first type of promotion information belongs, 2 promotion positions are matched to the promotion information source to which the second type of promotion information belongs, and 5 promotion positions are matched to the promotion information source to which the third type of promotion information belongs, then the exposure ratios corresponding to each promotion information source are 0.3, 0.2 and 0.5. When the promotion positions are matched this time, the exposure weight of the promotion information belonging to each promotion information source can adopt the exposure ratio of the promotion information source to which it belonged last time. By analogy, the next time the promotion position is matched, the exposure weight of the promotion information of each promotion information source can be based on the exposure ratio of the promotion information source to which it belongs at that time.

[0109] In this embodiment, the bidding score is also related to the exposure weight, and the exposure weight is related to the exposure ratio of the promotion information source after the previous promotion position allocation, which can achieve fair competition for the promotion position in the promotion page among the promotion information sources to which the candidate promotion information belongs. For example, only after the estimated exposure click rate and unit click value of the candidate information are improved, the exposure ratio of the promotion information source to which it belongs can be improved, so that it can obtain a higher exposure weight in the next promotion position competition, and achieve healthy competition among the promotion information sources.

[0110] In one embodiment, in order to distribute the promotional positions to different promotional information sources more evenly and display the promotional information from different promotional information sources, the server can obtain the initial exposure weight corresponding to each promotional information source. The initial exposure weight of each promotional information source enables the promotional positions to be evenly distributed to different promotional information sources in the initial stage of promotional position matching.

[0111] In one embodiment, the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs also includes: obtaining a preset exposure weight; iteratively adjusting the exposure weight of the promotion information source to which the candidate promotion information belongs starting from the preset exposure weight, and after matching the promotion position according to the exposure weight of the current iteration, obtaining the exposure share of the promotion information source to which the candidate promotion information belongs, and stopping the iterative adjustment when the exposure share meets the balanced exposure condition, thereby obtaining the initial exposure weight of the promotion information source to which the candidate promotion information belongs.

[0112] Specifically, the preset exposure weight may be the exposure weight of each promotion information source that is randomly initialized. The sum of the preset exposure weights corresponding to each promotion information source is 1. The server may calculate the bidding score of each historical candidate promotion information from the preset exposure weight, allocate promotion positions based on the bidding score, and then count the exposure share of each promotion information source under the current exposure weight. If the exposure share does not meet the balanced exposure condition, the preset exposure weight is adjusted with a preset gradient until the exposure share of each promotion information source meets the balanced exposure condition, and the adjusted preset exposure weight is used as the initial exposure weight of the promotion information source.

[0113] like Figure 6 As shown, it is a flowchart of calculating the initial exposure weight corresponding to the promotion information source in one embodiment. The server can sample the historical promotion samples corresponding to the historical candidate promotion information. The historical promotion samples can be the promotion data when the candidate promotion information is displayed in other promotion pages. Under the preset exposure weight of each promotion information source initialized, the exposure proportion of each promotion information source is calculated until the exposure weight adjusted through iteration is relatively stable and the corresponding exposure proportion meets the balanced promotion condition, then the adjustment will be ended, and the adjusted exposure weight will be used as the initial exposure weight of each promotion information source.

[0114] In this embodiment, the exposure share is counted based on the preset exposure weights of the promotion data corresponding to the historical candidate promotion information, so as to adjust the exposure share according to the preset gradient, and obtain the initial exposure weights that can evenly distribute the promotion positions to different promotion information sources in the initial stage of promotion position matching.

[0115] like Figure 7 FIG. 1 is a flow chart of a method for promoting position matching in a specific embodiment. The method for promoting position matching specifically includes the following steps:

[0116] S702, determining a promotion position and candidate promotion information corresponding to the user identifier.

[0117] S704: Acquire user characteristics corresponding to the user identifier.

[0118] S706: Extract promotion information features of the candidate promotion information.

[0119] S708: Generate a feature vector of the candidate promotion information according to the user feature and the promotion information feature.

[0120] S710, determining an estimated exposure click rate of the candidate promotion information through an exposure click rate prediction model corresponding to the promotion information source to which the candidate promotion information belongs and according to a feature vector of the candidate promotion information.

[0121] S712, when the candidate promotion information belongs to the first type of promotion information, obtaining the bid value of a single click of the candidate promotion information; and determining the bid value as the unit click value of the candidate promotion information.

[0122] S714, when the candidate promotion information belongs to the second type of promotion information, the average revenue value of the application downloaded after the candidate promotion information is clicked within a specified time period is obtained; and the unit click value of the candidate promotion information is determined according to the average revenue value.

[0123] S716, when the candidate promotion information belongs to the third type of promotion information, the first number of clicks on the third type of promotion information in the first promotion page is obtained; the promotion position of the first promotion page is only used to display the third type of promotion information; the second number of clicks on the third type of promotion information and the revenue value of non-third type promotion information in the second promotion page are obtained; the second promotion page includes a promotion position for displaying the third type of promotion information and a promotion position for displaying non-third type promotion information; the obtained revenue value is divided by the difference between the first number of clicks and the second number of clicks to obtain the unit click value of the candidate promotion information.

[0124] S718, counting the exposure ratio of the promotion information source to which the candidate promotion information belongs after the previous promotion position matching.

[0125] S720, determining the exposure weight of the promotion information source to which the candidate promotion information belongs when performing promotion position matching at that time according to the statistical exposure ratio.

[0126] S722, determining a bidding score corresponding to the candidate promotion information according to the estimated exposure click rate, the unit click value of the candidate promotion information, and the exposure weight of the promotion information source to which the candidate promotion information belongs.

[0127] S724, selecting promotion information whose corresponding bidding scores meet the promotion conditions from the candidate promotion information.

[0128] S726, matching the promotion position with the promotion information source to which the filtered promotion information belongs.

[0129] S728, receiving promotion information from the matched promotion information source; the received promotion information complies with the bidding ranking method of the matched promotion information source.

[0130] S730, pushing the received promotion information, where the pushed promotion information is used to be displayed in the matched promotion position.

[0131] The above-mentioned promotion position matching method matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click rate of the candidate promotion information, and the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0132] Figure 7 FIG. 1 is a flow chart of a method for generalizing a bit matching method in an embodiment. It should be understood that although Figure 7 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 7 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0133] like Figure 8 As shown, in one embodiment, a promotion information display method is provided. This embodiment mainly applies this method to the above Figure 1 The terminal 110 in FIG. Figure 8 , the promotion information display method specifically includes the following steps:

[0134] S802: Determine a promotion location corresponding to the user identifier.

[0135] Among them, the promotion slot is a slot on the Internet media used to display promotion information. The promotion slot can be an advertising slot on the App (Application, computer application) used to download applications for displaying various applications. For example, the recommendation list of "recommended applications" is displayed in the home page tab of the App for downloading applications. The promotion slot can also be an advertising slot in a social application, a scrolling advertisement bar on a portal website, a contextual recommendation label bar, a search result list in a search engine, and so on.

[0136] In one embodiment, after detecting the user's exposure behavior, the terminal can obtain the promotion position to be allocated corresponding to the user ID. The user ID and the promotion position to be allocated corresponding to the user ID are sent to the server, and after receiving the user ID, the server can obtain the promotion position to be allocated and candidate promotion information corresponding to the user ID.

[0137] The user's exposure behavior may be, for example, the user logging into an App used to download applications, or opening the App; it may also be the user browsing the social circle of a social application; it may also be the user opening a web page of a portal website, entering search keywords in a search engine, and so on.

[0138] S804, receiving promotion information that matches the promotion position; the promotion information comes from a promotion information source that matches the promotion position; the promotion information source is the promotion information source to which the candidate promotion information corresponding to the user identifier and whose corresponding bidding score meets the promotion conditions belongs; the bidding score is determined based on the estimated exposure click rate of the candidate promotion information and the unit click value of the candidate promotion information; the estimated exposure click rate is generated based on the user characteristics corresponding to the user identifier.

[0139] Among them, promotion information is the online communication medium of the promotion object. The promotion information can be text, pictures, keywords, rankings or videos. The computer device can use text description, picture display or video presentation to show the object to be promoted to the user. The promotion object can be, for example, an application to be downloaded, a product to be purchased, etc.

[0140] Candidate promotion information is a collection of promotion information corresponding to a user identifier. The types and quantities of candidate promotion information corresponding to different user identifiers may be different. In one embodiment, the candidate promotion information belongs to different promotion information sources. The promotion information source is the party that demands promotion positions. There may be multiple types of promotion information sources. For example, the promotion information source includes advertising parties, game parties, basic software parties, etc.

[0141] The Predicted Click Through Rate (PCTR) is the probability that the candidate promotion information will be clicked after being presented to the user through the promotion position corresponding to the user ID.

[0142] Cost Per Click (CPC) is the value that a candidate promotional message can bring when it is displayed in the promotional position corresponding to the user ID and clicked by the user once. The value of the promotional position can be quantified by the product of the number of clicks and the CPC, that is:

[0143] income = n × cpc, where n represents the number of times the promotional information displayed in the promotional position is clicked, cpc represents the unit click value of the displayed promotional information, and income represents the value of the promotional position.

[0144] The bidding score is a reference value for matching candidate promotion information with the promotion position corresponding to the user identifier. The bidding score of the candidate promotion information can be the product of the estimated exposure click rate and the unit click value of the candidate promotion information, or it can be the linear weighted sum of the estimated exposure click rate and the unit click value calculated under their respective weighting coefficients. When matching promotion positions based on the bidding score, the user's demand for promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0145] Specifically, after obtaining the promotion position to be allocated and the candidate promotion information corresponding to the user identifier, the bidding score of each candidate promotion information is calculated, the promotion information source matching the promotion position is determined based on the bidding score of each candidate provided information, and the promotion information from the matched promotion information source is received.

[0146] S806, displaying the matched promotion information in the promotion position.

[0147] In one embodiment, the steps for determining the estimated exposure click rate of candidate promotion information include: obtaining user characteristics corresponding to the user identifier; extracting promotion information characteristics of the candidate promotion information; generating a feature vector of the candidate promotion information based on the user characteristics and the promotion information characteristics; and determining the estimated exposure click rate of the candidate promotion information through an exposure click rate prediction model corresponding to the promotion information source to which the candidate promotion information belongs and based on the feature vector of the candidate promotion information.

[0148] In one embodiment, the training steps of the exposure click rate prediction model include: obtaining historical user characteristics, historical promotion information characteristics and sample labels corresponding to the historical promotion information belonging to the promotion information source; generating a historical feature vector corresponding to the promotion information belonging to the promotion information source based on the historical user characteristics and the historical promotion information characteristics; inputting the historical feature vector into the exposure click rate prediction model to obtain the historical exposure click rate corresponding to the historical feature vector; when it is determined that the training stop condition is not met based on the historical exposure click rate and the sample label, adjusting the model parameters of the exposure click rate prediction model and continuing training until the training stop condition is met.

[0149] In one embodiment, the step of determining the unit click value of the candidate promotion information includes: when the candidate promotion information belongs to the first type of promotion information, obtaining the bid value of a single click of the candidate promotion information; and determining the bid value as the unit click value of the candidate promotion information.

[0150] In one embodiment, the step of determining the unit click value of the candidate promotion information includes: when the candidate promotion information belongs to the second type of promotion information, obtaining the average revenue value of the application downloaded after the candidate promotion information is clicked within a specified time period; and determining the unit click value of the candidate promotion information based on the average revenue value.

[0151] In one embodiment, the step of determining the unit click value of the candidate promotional information includes: when the candidate promotional information belongs to the third type of promotional information, determining the estimated unit click value of the third type of promotional information; the estimated unit click value is determined when the total revenue value of the first promotional page and the total revenue value of the second promotional page are equal; the promotion position of the first promotional page is only used to display the third type of promotional information; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; the estimated unit click value is determined as the unit click value of the candidate promotional information.

[0152] In one embodiment, the step of determining the unit click value of the candidate promotional information includes: when the candidate promotional information belongs to the third type of promotional information, the unit click value of the candidate promotional information is the revenue value divided by the difference between the first number of clicks and the second number of clicks; the revenue value is the revenue value of the non-third type promotional information in the second promotional page; the first number of clicks is the number of clicks on the third type promotional information in the first promotional page; the second number of clicks is the number of clicks on the third type promotional information in the second promotional page; the promotion position of the first promotional page is only used to display the third type promotional information; the second promotional page includes a promotion position for displaying the third type promotional information and a promotion position for displaying non-third type promotional information.

[0153] In one embodiment, the bidding score is determined based on the estimated exposure click-through rate, the unit click value and the exposure weight of the promotion information source to which the candidate promotion information belongs; the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs includes: counting the exposure share of the promotion information source to which the candidate promotion information belongs after the previous promotion position matching; and determining the exposure weight of the promotion information source to which the candidate promotion information belongs when the promotion position matching is performed this time based on the counted exposure share.

[0154] In one embodiment, the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs also includes: obtaining a preset exposure weight; iteratively adjusting the exposure weight of the promotion information source to which the candidate promotion information belongs starting from the preset exposure weight, and after matching the promotion position according to the exposure weight of the current iteration, obtaining the exposure share of the promotion information source to which the candidate promotion information belongs, and stopping the iterative adjustment when the exposure share meets the balanced exposure condition, thereby obtaining the initial exposure weight of the promotion information source to which the candidate promotion information belongs.

[0155] The above-mentioned promotion information display method matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click rate of the candidate promotion information, and the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0156] like Fig. 9 As shown, in one embodiment, a promotion position matching device 900 is provided, the device 900 includes: a determination module 902, an acquisition module 904, a bidding score determination module 906, a screening module 908 and a matching module 910, wherein:

[0157] The determination module 902 is used to determine the promotion position and candidate promotion information corresponding to the user identifier.

[0158] The acquisition module 904 is used to acquire the estimated exposure click rate of the candidate promotion information; the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier.

[0159] The bidding score determination module 906 is used to determine the bidding score corresponding to the candidate promotion information according to the estimated exposure click rate and the unit click value of the candidate promotion information.

[0160] The screening module 908 is used to screen the promotion information whose corresponding bidding scores meet the promotion conditions from the candidate promotion information.

[0161] The matching module 910 is used to match the promotion position with the promotion information source to which the filtered promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source.

[0162] In one embodiment, Fig.10 As shown, the acquisition module 904 specifically includes a user feature acquisition module 1002, a promotion information feature extraction module 1004, a feature vector generation module 1006 and an estimated exposure click rate prediction module 1008. Among them, the user feature acquisition module 1002 is used to obtain user features corresponding to the user identifier; the promotion information feature extraction module 1004 is used to extract the promotion information features of the candidate promotion information; the feature vector generation module 1006 is used to generate a feature vector of the candidate promotion information according to the user features and the promotion information features; the estimated exposure click rate prediction module 1008 is used to determine the estimated exposure click rate of the candidate promotion information through the exposure click rate prediction model corresponding to the promotion information source to which the candidate promotion information belongs and according to the feature vector of the candidate promotion information.

[0163] In one embodiment, the exposure click rate prediction model training step in the estimated exposure click rate prediction module 1008 includes: obtaining historical user characteristics, historical promotion information characteristics and sample labels corresponding to the historical promotion information belonging to the promotion information source; generating a historical feature vector corresponding to the promotion information belonging to the promotion information source based on the historical user characteristics and the historical promotion information characteristics; inputting the historical feature vector into the exposure click rate prediction model to obtain the historical exposure click rate corresponding to the historical feature vector; when it is determined that the training stop condition is not met based on the historical exposure click rate and the sample label, adjusting the model parameters of the exposure click rate prediction model and continuing training until the training stop condition is met.

[0164] In one embodiment, the acquisition module 904 is further used to acquire the bid value of a single click of the candidate promotion information when the candidate promotion information belongs to the first type of promotion information; and determine the bid value as the unit click value of the candidate promotion information.

[0165] In one embodiment, when the candidate promotion information belongs to the second type of promotion information, the acquisition module 904 is also used to obtain the average revenue value of the application downloaded after the candidate promotion information is clicked within a specified period of time; and determine the unit click value of the candidate promotion information based on the average revenue value.

[0166] In one embodiment, the acquisition module 904 is also used to determine the estimated unit click value of the third type of promotional information when the candidate promotional information belongs to the third type of promotional information; the estimated unit click value is determined when the total revenue value of the first promotional page and the total revenue value of the second promotional page are equal; the promotion position of the first promotional page is only used to display the third type of promotional information; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; the estimated unit click value is determined as the unit click value of the candidate promotional information.

[0167] In one embodiment, the acquisition module 904 is also used to obtain the first number of clicks on the third type of promotional information in the first promotional page when the candidate promotional information belongs to the third type of promotional information; the promotion position of the first promotional page is only used to display the third type of promotional information; the second number of clicks on the third type of promotional information and the revenue value of non-third type promotional information in the second promotional page are obtained; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; the obtained revenue value is divided by the difference between the first number of clicks and the second number of clicks to obtain the unit click value of the candidate promotional information.

[0168] In one embodiment, the bidding score is determined based on the estimated exposure click-through rate, the unit click value and the exposure weight of the promotion information source to which the candidate promotion information belongs; the bidding score determination module 906 is also used to count the exposure share of the promotion information source to which the candidate promotion information belongs after the previous promotion position matching; based on the statistical exposure share, determine the exposure weight of the promotion information source to which the candidate promotion information belongs when the promotion position matching is performed this time.

[0169] In one embodiment, the bidding score determination module 906 is also used to obtain a preset exposure weight; iteratively adjust the exposure weight of the promotion information source to which the candidate promotion information belongs from the preset exposure weight, and after matching the promotion position according to the exposure weight of the current iteration, obtain the exposure share of the promotion information source to which the candidate promotion information belongs, and stop the iterative adjustment when the exposure share meets the balanced exposure condition, so as to obtain the initial exposure weight of the promotion information source to which the candidate promotion information belongs.

[0170] In one embodiment, Fig.11 As shown, the promotion position matching device 900 also includes a promotion information receiving module 1102 and a promotion information pushing module 1104, wherein: the promotion information receiving module 1102 is used to receive promotion information from the matched promotion information source; the received promotion information complies with the bidding sorting method of the matched promotion information source; the promotion information pushing module 1104 is used to push the received promotion information, and the pushed promotion information is used to be displayed in the matched promotion position.

[0171] The above-mentioned promotion position matching device 900 matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click rate of the candidate promotion information, and the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all the candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0172] like Fig.12 As shown, in one embodiment, a promotion information display device 1200 is provided, and the device 1200 includes: a promotion position determination module 1202, a promotion information receiving module 1204 and a display module 1206, wherein:

[0173] The promotion position determination module 1202 is used to determine the promotion position corresponding to the user identifier.

[0174] The promotion information receiving module 1204 is used to receive promotion information matching the promotion position; the promotion information comes from the promotion information source matching the promotion position; the promotion information source is the promotion information source to which the candidate promotion information corresponding to the user identifier and whose corresponding bidding score meets the promotion conditions belongs; the bidding score is determined based on the estimated exposure click-through rate of the candidate promotion information and the unit click value of the candidate promotion information; the estimated exposure click-through rate is generated based on the user characteristics corresponding to the user identifier.

[0175] The display module 1206 is used to display the matched promotion information in the promotion position.

[0176] In one embodiment, when the candidate promotional information belongs to the third type of promotional information, the unit click value of the candidate promotional information is the revenue value divided by the difference between the first number of clicks and the second number of clicks; the revenue value is the revenue value of non-third type promotional information in the second promotional page; the first number of clicks is the number of clicks on the third type promotional information in the first promotional page; the second number of clicks is the number of clicks on the third type promotional information in the second promotional page; the promotion position of the first promotional page is only used to display the third type of promotional information; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information.

[0177] The above-mentioned promotion information display device 1200 matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click rate of the candidate promotion information, and the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0178] Fig.13 The internal structure diagram of a computer device in one embodiment is shown. The computer device may specifically be Figure 1 The server 120 in FIG. Fig.13 As shown, the computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program, which, when executed by the processor, enables the processor to implement the generalized bit matching method. The internal memory may also store a computer program, which, when executed by the processor, enables the processor to execute the generalized bit matching method.

[0179] In one embodiment, the promotion position matching device provided by the present application can be implemented in the form of a computer program. The computer program can be Fig.13 The computer device shown in the figure is run. The memory of the computer device can store various program modules constituting the promotion position matching device, such as: Fig. 9The determination module 902, acquisition module 904, bidding score determination module 906, screening module 908 and matching module 910 are shown. The computer program composed of various program modules enables the processor to execute the steps of the promotion position matching method of each embodiment of the present application described in this specification.

[0180] For example, Fig.13 The computer device shown can be Fig. 9 The determination module 902 in the promotion position matching device shown executes step S202. The computer device may execute step S204 through the acquisition module 904. The computer device may execute step S206 through the bidding score determination module 906. The computer device may execute step S208 through the screening module 908. The computer device may execute step S210 through the matching module 910.

[0181] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the following steps: determining a promotion position and candidate promotion information corresponding to a user identifier; obtaining an estimated exposure click-through rate of the candidate promotion information; generating an estimated exposure click-through rate based on user characteristics corresponding to the user identifier; determining a bidding score corresponding to the candidate promotion information based on the estimated exposure click-through rate and a unit click value of the candidate promotion information; screening promotion information whose corresponding bidding scores meet promotion conditions from the candidate promotion information; matching the promotion position with a promotion information source to which the screened promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source.

[0182] In one embodiment, when the computer program is executed by the processor to obtain the step of obtaining the estimated exposure click rate of the candidate promotion information, the processor specifically performs the following steps: obtaining user characteristics corresponding to the user identifier; extracting the promotion information characteristics of the candidate promotion information; generating a feature vector of the candidate promotion information based on the user characteristics and the promotion information characteristics; and determining the estimated exposure click rate of the candidate promotion information through an exposure click rate prediction model corresponding to the promotion information source to which the candidate promotion information belongs and based on the feature vector of the candidate promotion information.

[0183] In one embodiment, the training steps of the exposure click rate prediction model include: obtaining historical user characteristics, historical promotion information characteristics and sample labels corresponding to the historical promotion information belonging to the promotion information source; generating a historical feature vector corresponding to the promotion information belonging to the promotion information source based on the historical user characteristics and the historical promotion information characteristics; inputting the historical feature vector into the exposure click rate prediction model to obtain the historical exposure click rate corresponding to the historical feature vector; when it is determined that the training stop condition is not met based on the historical exposure click rate and the sample label, adjusting the model parameters of the exposure click rate prediction model and continuing training until the training stop condition is met.

[0184] In one embodiment, the step of determining the unit click value of the candidate promotion information includes: when the candidate promotion information belongs to the first type of promotion information, obtaining the bid value of a single click of the candidate promotion information; and determining the bid value as the unit click value of the candidate promotion information.

[0185] In one embodiment, the step of determining the unit click value of the candidate promotion information includes: when the candidate promotion information belongs to the second type of promotion information, obtaining the average revenue value of the application downloaded after the candidate promotion information is clicked within a specified time period; and determining the unit click value of the candidate promotion information based on the average revenue value.

[0186] In one embodiment, the step of determining the unit click value of the candidate promotional information includes: when the candidate promotional information belongs to the third type of promotional information, determining the estimated unit click value of the third type of promotional information; the estimated unit click value is determined when the total revenue value of the first promotional page and the total revenue value of the second promotional page are equal; the promotion position of the first promotional page is only used to display the third type of promotional information; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; the estimated unit click value is determined as the unit click value of the candidate promotional information.

[0187] In one embodiment, the step of determining the unit click value of the candidate promotional information includes: when the candidate promotional information belongs to the third type of promotional information, obtaining the first number of clicks of the third type of promotional information in the first promotional page; the promotion position of the first promotional page is only used to display the third type of promotional information; obtaining the second number of clicks of the third type of promotional information and the revenue value of non-third type promotional information in the second promotional page; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; dividing the obtained revenue value by the difference between the first number of clicks and the second number of clicks to obtain the unit click value of the candidate promotional information.

[0188] In one embodiment, the bidding score is determined based on the estimated exposure click-through rate, the unit click value and the exposure weight of the promotion information source to which the candidate promotion information belongs; the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs includes: counting the exposure share of the promotion information source to which the candidate promotion information belongs after the previous promotion position matching; and determining the exposure weight of the promotion information source to which the candidate promotion information belongs when the promotion position matching is performed this time based on the counted exposure share.

[0189] In one embodiment, the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs also includes: obtaining a preset exposure weight; iteratively adjusting the exposure weight of the promotion information source to which the candidate promotion information belongs starting from the preset exposure weight, and after matching the promotion position according to the exposure weight of the current iteration, obtaining the exposure share of the promotion information source to which the candidate promotion information belongs, and stopping the iterative adjustment when the exposure share meets the balanced exposure condition, thereby obtaining the initial exposure weight of the promotion information source to which the candidate promotion information belongs.

[0190] In one embodiment, when the computer program is executed by a processor, the processor further performs the following steps: receiving promotional information from a matched promotional information source; the received promotional information complies with the bidding sorting method of the matched promotional information source; and pushing the received promotional information, and the pushed promotional information is used for display in the matched promotional position.

[0191] The above-mentioned computer device matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click rate of the candidate promotion information, and the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all the candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0192] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the processor executes the following steps: determining a promotion position and candidate promotion information corresponding to a user identifier; obtaining an estimated exposure click-through rate of the candidate promotion information; generating an estimated exposure click-through rate based on user characteristics corresponding to the user identifier; determining a bidding score corresponding to the candidate promotion information based on the estimated exposure click-through rate and the unit click value of the candidate promotion information; screening promotion information whose corresponding bidding scores meet the promotion conditions from the candidate promotion information; matching the promotion position with a promotion information source to which the screened promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source.

[0193] In one embodiment, when the computer program is executed by the processor to obtain the step of obtaining the estimated exposure click rate of the candidate promotion information, the processor specifically performs the following steps: obtaining user characteristics corresponding to the user identifier; extracting the promotion information characteristics of the candidate promotion information; generating a feature vector of the candidate promotion information based on the user characteristics and the promotion information characteristics; and determining the estimated exposure click rate of the candidate promotion information through an exposure click rate prediction model corresponding to the promotion information source to which the candidate promotion information belongs and based on the feature vector of the candidate promotion information.

[0194] In one embodiment, the training steps of the exposure click rate prediction model include: obtaining historical user characteristics, historical promotion information characteristics and sample labels corresponding to the historical promotion information belonging to the promotion information source; generating a historical feature vector corresponding to the promotion information belonging to the promotion information source based on the historical user characteristics and the historical promotion information characteristics; inputting the historical feature vector into the exposure click rate prediction model to obtain the historical exposure click rate corresponding to the historical feature vector; when it is determined that the training stop condition is not met based on the historical exposure click rate and the sample label, adjusting the model parameters of the exposure click rate prediction model and continuing training until the training stop condition is met.

[0195] In one embodiment, the step of determining the unit click value of the candidate promotion information includes: when the candidate promotion information belongs to the first type of promotion information, obtaining the bid value of a single click of the candidate promotion information; and determining the bid value as the unit click value of the candidate promotion information.

[0196] In one embodiment, the step of determining the unit click value of the candidate promotion information includes: when the candidate promotion information belongs to the second type of promotion information, obtaining the average revenue value of the application downloaded after the candidate promotion information is clicked within a specified time period; and determining the unit click value of the candidate promotion information based on the average revenue value.

[0197] In one embodiment, the step of determining the unit click value of the candidate promotional information includes: when the candidate promotional information belongs to the third type of promotional information, determining the estimated unit click value of the third type of promotional information; the estimated unit click value is determined when the total revenue value of the first promotional page and the total revenue value of the second promotional page are equal; the promotion position of the first promotional page is only used to display the third type of promotional information; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; the estimated unit click value is determined as the unit click value of the candidate promotional information.

[0198] In one embodiment, the step of determining the unit click value of the candidate promotional information includes: when the candidate promotional information belongs to the third type of promotional information, obtaining the first number of clicks of the third type of promotional information in the first promotional page; the promotion position of the first promotional page is only used to display the third type of promotional information; obtaining the second number of clicks of the third type of promotional information and the revenue value of non-third type promotional information in the second promotional page; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information; dividing the obtained revenue value by the difference between the first number of clicks and the second number of clicks to obtain the unit click value of the candidate promotional information.

[0199] In one embodiment, the bidding score is determined based on the estimated exposure click-through rate, the unit click value and the exposure weight of the promotion information source to which the candidate promotion information belongs; the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs includes: counting the exposure share of the promotion information source to which the candidate promotion information belongs after the previous promotion position matching; and determining the exposure weight of the promotion information source to which the candidate promotion information belongs when the promotion position matching is performed this time based on the counted exposure share.

[0200] In one embodiment, the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs also includes: obtaining a preset exposure weight; iteratively adjusting the exposure weight of the promotion information source to which the candidate promotion information belongs starting from the preset exposure weight, and after matching the promotion position according to the exposure weight of the current iteration, obtaining the exposure share of the promotion information source to which the candidate promotion information belongs, and stopping the iterative adjustment when the exposure share meets the balanced exposure condition, thereby obtaining the initial exposure weight of the promotion information source to which the candidate promotion information belongs.

[0201] In one embodiment, when the computer program is executed by a processor, the processor further performs the following steps: receiving promotional information from a matched promotional information source; the received promotional information complies with the bidding sorting method of the matched promotional information source; and pushing the received promotional information, and the pushed promotional information is used for display in the matched promotional position.

[0202] The above-mentioned computer-readable storage medium matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click-through rate of the candidate promotion information, and the estimated exposure click-through rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click-through rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0203] Fig.14 The internal structure diagram of a computer device in one embodiment is shown. The computer device may specifically be Figure 1 The terminal 110 in FIG. Fig.14 As shown, the computer device includes a processor, a memory, a network interface, an input device and a display screen connected through a system bus. Among them, the memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program. When the computer program is executed by the processor, the processor can implement the promotion information display method. The internal memory may also store a computer program. When the computer program is executed by the processor, the processor can execute the promotion information display method. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0204] Those skilled in the art will understand that Fig.13 and 14 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0205] In one embodiment, the promotion information display device provided by the present application can be implemented in the form of a computer program. The computer program can be Fig.14 The computer device memory may store various program modules constituting the promotion information display device, for example, Fig.12The promotion position determination module 1202, the promotion information receiving module 1204 and the display module 1206 are shown. The computer program composed of various program modules enables the processor to execute the steps of the promotion information display method of each embodiment of the present application described in this specification.

[0206] For example, Fig.14 The computer device shown can be Fig.12 The determination module 1202 in the promotion information display device shown executes step S1102. The computer device may execute step S1104 through the promotion information receiving module 1204. The computer device may execute step S1106 through the display module 1206.

[0207] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the following steps: determining a promotion position corresponding to a user identifier; receiving promotion information matching the promotion position; the promotion information comes from a promotion information source matching the promotion position; the promotion information source is a promotion information source to which candidate promotion information corresponding to the user identifier and whose corresponding bidding score meets the promotion conditions belongs; the bidding score is determined based on an estimated exposure click-through rate of the candidate promotion information and a unit click value of the candidate promotion information; the estimated exposure click-through rate is generated based on user characteristics corresponding to the user identifier; and the matched promotion information is displayed in the promotion position.

[0208] In one embodiment, when the candidate promotional information belongs to the third type of promotional information, the unit click value of the candidate promotional information is the revenue value divided by the difference between the first number of clicks and the second number of clicks; the revenue value is the revenue value of non-third type promotional information in the second promotional page; the first number of clicks is the number of clicks on the third type promotional information in the first promotional page; the second number of clicks is the number of clicks on the third type promotional information in the second promotional page; the promotion position of the first promotional page is only used to display the third type of promotional information; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information.

[0209] The above-mentioned computer device matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click rate of the candidate promotion information, and the estimated exposure click rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all the candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0210] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the processor executes the following steps: determining a promotion position corresponding to a user identifier; receiving promotion information matching the promotion position; the promotion information comes from a promotion information source matching the promotion position; the promotion information source is a promotion information source to which candidate promotion information corresponding to the user identifier and whose corresponding bidding score meets the promotion conditions belongs; the bidding score is determined based on an estimated exposure click-through rate of the candidate promotion information and a unit click value of the candidate promotion information; the estimated exposure click-through rate is generated based on user characteristics corresponding to the user identifier; and the matched promotion information is displayed in the promotion position.

[0211] In one embodiment, when the candidate promotional information belongs to the third type of promotional information, the unit click value of the candidate promotional information is the revenue value divided by the difference between the first number of clicks and the second number of clicks; the revenue value is the revenue value of non-third type promotional information in the second promotional page; the first number of clicks is the number of clicks on the third type promotional information in the first promotional page; the second number of clicks is the number of clicks on the third type promotional information in the second promotional page; the promotion position of the first promotional page is only used to display the third type of promotional information; the second promotional page includes a promotion position for displaying the third type of promotional information and a promotion position for displaying non-third type promotional information.

[0212] The above-mentioned computer-readable storage medium matches the promotion position with the promotion information source based on the bidding score of the candidate promotion information belonging to different promotion information sources for the promotion position. Among them, the bidding score is determined according to the estimated exposure click-through rate of the candidate promotion information, and the estimated exposure click-through rate is generated according to the user characteristics corresponding to the user identifier, so that the estimated exposure click-through rate can fully reflect the matching situation between the candidate promotion information and the user. The unit click value of the candidate promotion information can reflect the value advantage of the promotion information among all candidate information. In this way, when matching the promotion position based on the bidding score, the user's demand for the promotion information in the promotion information source and the value advantage of the promotion information in the promotion information source can be fully considered, so that the promotion position resources can be fully utilized and waste can be avoided.

[0213] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0214] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A promotion position matching method, comprising: A user identifier and a promotion location corresponding to the user identifier sent by the receiving terminal after detecting the exposure behavior of the user, wherein the promotion location is used to display promotion information; Determine candidate promotion information corresponding to the user identifier, the promotion information sources of the candidate promotion information include an advertisement party, a game party and a basic software party, the candidate promotion information includes a first type of promotion information from an advertisement party, a second type of promotion information from a game party and a third type of promotion information from a basic software party, and the candidate promotion information belonging to different promotion information sources is used to promote different types of applications; Obtaining an estimated exposure click rate of the candidate promotion information; the estimated exposure click rate is generated according to a user feature corresponding to the user identifier; When the candidate promotion information belongs to the first type of promotion information from the advertiser, obtaining the bid value paid by the advertiser for a single click on the candidate promotion information, and determining the bid value as the unit click value of the candidate promotion information; When the candidate promotion information belongs to the second type of promotion information from the game party, obtaining the average revenue value of the game application downloaded after the candidate promotion information is clicked within a specified time period, and determining the obtained average revenue value as the unit click value of the candidate promotion information; When the candidate promotion information belongs to the third type of promotion information from the basic software party, then when the total revenue value of the first promotion page and the total revenue value of the second promotion page are equal, obtain the first number of clicks on the third type of promotion information in the first promotion page, obtain the second number of clicks on the third type of promotion information in the second promotion page and the revenue value of non-third type promotion information, the promotion position of the first promotion page is only used to display the third type of promotion information, the second promotion page includes the promotion position for displaying the third type of promotion information and the promotion position for displaying non-third type promotion information, divide the obtained revenue value by the difference between the first number of clicks and the second number of clicks, and obtain the unit click value of the candidate promotion information; Determining a bidding score corresponding to the candidate promotion information according to the estimated exposure click rate and the unit click value of the candidate promotion information; Filtering promotion information whose corresponding bidding scores meet promotion conditions from the candidate promotion information; The promotion position is matched with the promotion information source to which the filtered promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source.

2. The method according to claim 1, characterized in that: The obtaining of the estimated exposure click rate of the candidate promotion information includes: Acquire user characteristics corresponding to the user identifier; Extracting promotion information features of the candidate promotion information; Generating a feature vector of the candidate promotion information according to the user feature and the promotion information feature; The estimated exposure click rate of the candidate promotion information is determined by using an exposure click rate prediction model corresponding to the promotion information source to which the candidate promotion information belongs and according to the feature vector of the candidate promotion information.

3. The method according to claim 2, characterized in that The training steps of the exposure click rate prediction model include: Acquire historical user features, historical promotion information features and sample labels corresponding to the historical promotion information belonging to the promotion information source; Generate a historical feature vector corresponding to the promotion information belonging to the promotion information source according to the historical user features and the historical promotion information features; Inputting the historical feature vector into the exposure click rate prediction model to obtain a historical exposure click rate corresponding to the historical feature vector; When it is determined that the training stop condition is not met based on the historical exposure click-through rate and the sample label, the model parameters of the exposure click-through rate prediction model are adjusted and training is continued until the training stop condition is met.

4. The method according to claim 1, characterized in that The bidding score is determined according to the estimated exposure click rate, the unit click value and the exposure weight of the promotion information source to which the candidate promotion information belongs; the step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs includes: Count the exposure ratio of the promotion information source to which the candidate promotion information belongs after the previous promotion position matching; According to the statistical exposure ratio, the exposure weight of the promotion information source to which the candidate promotion information belongs when matching the promotion position is determined.

5. The method according to claim 4, characterized in that The step of determining the exposure weight of the promotion information source to which the candidate promotion information belongs also includes: Get the preset exposure weight; The exposure weight of the promotion information source to which the candidate promotion information belongs is iteratively adjusted starting from the preset exposure weight, and after matching the promotion position according to the exposure weight of the current iteration, the exposure share of the promotion information source to which the candidate promotion information belongs is obtained, and the iterative adjustment is stopped until the exposure share meets the balanced exposure condition, so as to obtain the initial exposure weight of the promotion information source to which the candidate promotion information belongs.

6. The method according to any one of claims 1 to 5, characterized in that The method further comprises: Receiving promotional information from the matched promotional information source; the received promotional information complies with the bidding ranking method of the matched promotional information source; The received promotion information is pushed, and the pushed promotion information is used to be displayed in the matched promotion position.

7. A promotion position matching device, characterized in that: The device comprises: A determination module is used to receive a user identifier and a promotion position corresponding to the user identifier sent by the terminal after detecting the user's exposure behavior, wherein the promotion position is used to display promotion information; determine candidate promotion information corresponding to the user identifier, wherein the promotion information sources of the candidate promotion information include an advertising party, a game party, and a basic software party, wherein the candidate promotion information includes a first type of promotion information from an advertising party, a second type of promotion information from a game party, and a third type of promotion information from a basic software party, and the candidate promotion information belonging to different promotion information sources is used to promote different types of applications; An acquisition module, used to acquire an estimated exposure click rate of the candidate promotion information; the estimated exposure click rate is generated according to a user feature corresponding to the user identifier; The acquisition module is further configured to, when the candidate promotion information belongs to the first type of promotion information from the advertiser, acquire the bid value paid by the advertiser for a single click on the candidate promotion information, and determine the bid value as the unit click value of the candidate promotion information; The acquisition module is further configured to, when the candidate promotion information belongs to the second type of promotion information from the game party, acquire the average revenue value of the game application downloaded after the candidate promotion information is clicked within a specified time period, and determine the acquired average revenue value as the unit click value of the candidate promotion information; The acquisition module is further used for, when the candidate promotion information belongs to the third type of promotion information from the basic software party, obtaining a first number of clicks on the third type of promotion information in the first promotion page and a second number of clicks on the third type of promotion information in the second promotion page, and obtaining a second number of clicks on the third type of promotion information and a revenue value of non-third type promotion information in the second promotion page, when the total revenue value of the first promotion page and the total revenue value of the second promotion page are equal, the promotion position of the first promotion page is only used to display the third type of promotion information, the second promotion page includes a promotion position for displaying the third type of promotion information and a promotion position for displaying non-third type promotion information, and dividing the obtained revenue value by the difference between the first number of clicks and the second number of clicks to obtain a unit click value of the candidate promotion information; A bidding score determination module, used to determine the bidding score corresponding to the candidate promotion information according to the estimated exposure click rate and the unit click value of the candidate promotion information; A screening module, used to screen the promotion information whose corresponding bidding scores meet the promotion conditions from the candidate promotion information; The matching module is used to match the promotion position with the promotion information source to which the screened promotion information belongs; the promotion position is used to display the promotion information from the matched promotion information source.

8. The device according to claim 7, characterized in that The acquisition module specifically includes: A user feature acquisition module, used to acquire the user feature corresponding to the user identifier; A promotion information feature extraction module, used to extract promotion information features of the candidate promotion information; A feature vector generating module, used for generating a feature vector of the candidate promotion information according to the user feature and the promotion information feature; The estimated exposure click rate prediction module is used to determine the estimated exposure click rate of the candidate promotion information through the exposure click rate prediction model corresponding to the promotion information source to which the candidate promotion information belongs and according to the feature vector of the candidate promotion information.

9. The device according to claim 8, characterized in that The device also includes: A training module is used to obtain historical user features, historical promotion information features and sample labels corresponding to the historical promotion information belonging to the promotion information source; generate a historical feature vector corresponding to the promotion information belonging to the promotion information source based on the historical user features and historical promotion information features; input the historical feature vector into the exposure click-through rate prediction model to obtain a historical exposure click-through rate corresponding to the historical feature vector; when it is determined that the training stop condition is not met based on the historical exposure click-through rate and the sample label, adjust the model parameters of the exposure click-through rate prediction model and continue training until the training stop condition is met.

10. The device according to claim 7, characterized in that The bidding score is determined based on the estimated exposure click rate, the unit click value and the exposure weight of the promotion information source to which the candidate promotion information belongs. The bidding score determination module is also used to count the exposure ratio of the promotion information source to which the candidate promotion information belongs after the previous promotion position matching; based on the counted exposure ratio, determine the exposure weight of the promotion information source to which the candidate promotion information belongs when the promotion position matching is performed this time.

11. The device according to claim 10, characterized in that The bidding score determination module is also used to obtain a preset exposure weight; iteratively adjust the exposure weight of the promotion information source to which the candidate promotion information belongs from the preset exposure weight, and after matching the promotion position according to the exposure weight of the current iteration, obtain the exposure share of the promotion information source to which the candidate promotion information belongs, and stop the iterative adjustment when the exposure share meets the balanced exposure condition, so as to obtain the initial exposure weight of the promotion information source to which the candidate promotion information belongs.

12. The device according to any one of claims 7 to 11, characterized in that The device also includes: A promotion information receiving module, used to receive promotion information from a matched promotion information source, where the received promotion information complies with the bidding ranking method of the matched promotion information source; The promotion information push module is used to push the received promotion information, and the pushed promotion information is used to be displayed in the matched promotion position.

13. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 6.

14. A computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 6.

Images