Advertisement inventory inquiry method, device, equipment and storage medium

Abstract

The embodiments of the present application disclose a method, device, equipment and storage medium for querying advertising inventory, wherein the method includes: determining the inventory corresponding to each advertising position on the target advertising delivery page; determining the theoretical available inventory corresponding to the target advertisement; correcting the theoretical available inventory corresponding to the target advertisement by the loss coefficient to obtain the actual available inventory corresponding to the target advertisement, wherein the loss coefficient can characterize the degree of influence of the advertising delivery system on the advertising exposure when the advertising delivery system adopts the deduplication rule for advertising delivery. The method takes into account the influence of the deduplication rule in the query process, and uses the loss coefficient to correct the theoretical available inventory, so as to ensure that the actual available inventory finally obtained is more accurate, reduce the occurrence of advertising absence, and ensure that the actual needs of the contract advertisers are met.

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to an advertising inventory inquiry method, device, equipment and storage medium. Background Art

[0002] At present, advertising delivery platforms provide contract advertising delivery services to advertisers. The so-called contract advertising refers to advertising with guaranteed price and quantity, which means that the advertising exposure volume is guaranteed to meet the advertising delivery requirements of the advertiser. Therefore, when selling traffic, the advertising delivery platform needs to know the current advertising inventory. At this time, it is necessary to determine the advertising inventory through the advertising inventory inquiry method. If the inquiry result is too high, it will cause the advertisement to be missed and fail to meet the actual needs of the contract advertiser.

[0003] The existing advertising inventory inquiry method assumes that each ad space is independent and unrelated to each other, and is not compatible with the ad deduplication rules. However, in actual applications, advertising is delivered based on the ad deduplication rules, which leads to higher inquiry results. Summary of the invention

[0004] The embodiments of the present application provide a method, apparatus, device and storage medium for querying advertising inventory. In the process of querying advertising inventory, advertising deduplication rules are comprehensively considered to ensure that the final query result is more accurate.

[0005] In view of this, the first aspect of the present application provides an advertising inventory inquiry method, the method comprising:

[0006] Determine the inventory corresponding to each ad slot on the target ad delivery page;

[0007] Determine the theoretical available inventory corresponding to the target advertisement to be delivered;

[0008] The theoretical available inventory corresponding to the target advertisement is corrected by a depreciation coefficient to obtain the actual available inventory corresponding to the target advertisement. The depreciation coefficient represents the degree of influence on the advertisement exposure when the advertisement delivery system adopts deduplication rules for advertisement delivery.

[0009] A second aspect of the present application provides an advertising inventory inquiry device, the device comprising:

[0010] The first determination module is used to determine the inventory corresponding to each advertising position on the target advertising delivery page;

[0011] The second determination module is used to determine the theoretical available inventory corresponding to the target advertisement to be delivered;

[0012] The correction module is used to correct the theoretical available inventory corresponding to the target advertisement by using a depreciation coefficient to obtain the actual available inventory corresponding to the target advertisement. The depreciation coefficient represents the degree of influence of the advertisement delivery system on the advertisement exposure when the deduplication rule is used for advertisement delivery.

[0013] A third aspect of the present application provides a device, the device comprising a processor and a memory:

[0014] The memory is used to store program code and transmit the program code to the processor;

[0015] The processor is used to execute the steps of the advertising inventory inquiry method as described in the first aspect according to the instructions in the program code.

[0016] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium is used to store program code, and the program code is used to execute the advertising inventory inquiry method described in the first aspect above.

[0017] A fifth aspect of the present application provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the advertising inventory inquiry method described in the first aspect.

[0018] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0019] The embodiment of the present application provides a method for querying advertising inventory, which takes into account the impact of deduplication rules on the query results of advertising inventory during the query process. Specifically, when querying the target advertisement, first determine the inventory corresponding to each ad slot on the target advertisement delivery page, and then determine the theoretical available inventory corresponding to the target advertisement based on the inventory corresponding to each ad slot and the targeting conditions of the target advertisement; then, correct the theoretical available inventory corresponding to the target advertisement by using the deduplication coefficient to obtain the actual available inventory corresponding to the target advertisement, wherein the deduplication coefficient can characterize the degree of influence of the advertising delivery system on the advertising exposure when the deduplication rules are used for advertising delivery. In this way, the theoretical available inventory is corrected by using the deduplication coefficient to obtain the actual available inventory, ensuring that the final determined actual available inventory is determined while considering the impact of the deduplication rules, that is, ensuring that the final determined query result is more accurate, reducing the occurrence of advertisement absence, and ensuring that the actual needs of the contract advertisers are met. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic diagram of the principle of deduplication for advertising pages;

[0021] Figure 2A schematic diagram of an application scenario of the advertising inventory inquiry method provided in an embodiment of the present application;

[0022] Figure 3 A flowchart of an advertising inventory inquiry method provided in an embodiment of the present application;

[0023] Figure 4 A schematic diagram of a method for deriving actual available inventory corresponding to a target advertisement provided in an embodiment of the present application;

[0024] Figure 5 A schematic diagram of the structure of a first advertising inventory inquiry device provided in an embodiment of the present application;

[0025] Figure 6 A schematic diagram of the structure of a second advertising inventory inquiry device provided in an embodiment of the present application;

[0026] Figure 7 A schematic diagram of the structure of a third advertising inventory inquiry device provided in an embodiment of the present application;

[0027] Figure 8 A schematic diagram of the structure of a fourth advertising inventory inquiry device provided in an embodiment of the present application;

[0028] Fig. 9 A schematic diagram of the structure of a fifth advertising inventory inquiry device provided in an embodiment of the present application;

[0029] Fig.10 A schematic diagram of the structure of a sixth advertising inventory inquiry device provided in an embodiment of the present application;

[0030] Fig.11 A schematic diagram of the structure of a server provided in an embodiment of the present application;

[0031] Fig.12 A schematic diagram of the structure of a terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0033] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0034] When an advertising platform provides contract advertising services to advertisers, it usually needs to inquire about the advertising inventory based on the targeting conditions of the contract advertising, obtain the remaining advertising inventory that meets the targeting conditions, and then set the scheduled exposure of the contract advertising based on the remaining advertising inventory, that is, set the available inventory corresponding to the contract advertising. If the available inventory determined by the inquiry is too high, it will cause the advertisement to be missing and fail to meet the needs of advertisers; if the available inventory determined by the inquiry is too low, it will result in the inability to fully sell the traffic, affecting the revenue of the advertising platform. It can be seen that accurately determining the available inventory corresponding to the advertisement is of great significance to both the advertising platform and the advertiser.

[0035] Nowadays, in order to further optimize the user experience, major advertising platforms have added deduplication rules in the process of advertising, that is, they require that the same advertisements cannot be delivered within a certain window range, and the advertisements need to be deduplicated within the window range. Figure 1 , Figure 1 This is a schematic diagram of advertising placement corresponding to the deduplication principle; Figure 1 As shown, assuming that the deduplication window range of a certain advertising delivery page is 3, the same advertisement cannot be delivered within a certain deduplication window range of the advertising delivery page, that is, the advertisements delivered in ad slot 1, ad slot 2, and ad slot 3 within a certain deduplication window range cannot be the same.

[0036] However, the existing advertising inventory query method assumes that each advertising position is independent and unrelated. Therefore, when querying advertising inventory, the existing technology usually does not consider the impact of deduplication rules on the advertising inventory query results, which leads to high query results and frequent advertising omissions.

[0037] In order to solve the problems existing in the above-mentioned prior art, an embodiment of the present application provides an advertising inventory inquiry method. When determining the advertising inventory inquiry result, the method corrects the theoretical available inventory corresponding to the advertisement with the help of a depreciation coefficient. The depreciation coefficient can characterize the degree of influence of the advertising delivery system on the advertising exposure when using deduplication rules for advertising delivery. In this way, the influence of deduplication rules on the inquiry result is considered in the process of determining the advertising inventory inquiry result, thereby ensuring that the final determined inquiry result is more accurate, reducing the occurrence of advertising omissions, and ensuring that the actual needs of advertisers are met.

[0038] It should be understood that the advertising inventory inquiry method provided in the embodiment of the present application can be applied to devices with data processing functions, such as terminal devices, servers, etc. Among them, the terminal device can specifically be a computer, a personal digital assistant (PDA), a smart phone, a tablet computer, etc.; the server can specifically be an application server or a Web server. In actual application deployment, the server can be an independent server or a cluster server.

[0039] In order to facilitate understanding of the technical solution provided in the embodiment of the present application, the advertising inventory inquiry method provided in the embodiment of the present application is introduced below by taking the server performing advertising inventory inquiry as an example.

[0040] See also Figure 2 , Figure 2 A schematic diagram of an application scenario of the advertising inventory inquiry method provided in the embodiment of the present application. The application scenario includes: a server 201, which is used to execute the advertising inventory inquiry method provided in the embodiment of the present application, and determine the actual available inventory corresponding to the target advertisement.

[0041] When an advertiser needs to deliver a target advertisement through an advertisement delivery platform, the advertisement delivery platform needs to use server 201 to query the advertisement inventory for the target advertisement to determine the actual available inventory corresponding to the target advertisement, that is, determine the exposure corresponding to the target advertisement.

[0042] When specifically inquiring about the advertising inventory, server 201 first determines the inventory corresponding to each ad slot on the target advertising delivery page; then, based on the inventory corresponding to each ad slot and the targeting conditions of the target advertisement, server 201 determines the theoretical available inventory corresponding to the target advertisement; further, server 201 uses the depreciation coefficient to correct the theoretical available inventory corresponding to the target advertisement, and obtains the actual available inventory corresponding to the target advertisement, wherein the depreciation coefficient can characterize the degree of influence of the advertising delivery system on the advertising exposure when using deduplication rules for advertising delivery.

[0043] In the process of querying the advertising inventory, server 201 uses the depreciation coefficient to correct the theoretical available inventory corresponding to the target advertisement, and obtains the actual available inventory corresponding to the target advertisement. In this way, the impact of the deduplication rule on the advertising inventory query result is comprehensively considered to ensure that the final query result is more accurate, reduce the occurrence of advertising shortages, and ensure that the actual needs of advertisers of the target advertisement are met.

[0044] It should be understood that the above Figure 2 The application scenario shown is only an example. In actual applications, the advertising inventory inquiry method provided in the embodiment of the present application can also be independently executed by the terminal device. No limitation is made to the application scenario of the advertising inventory inquiry method provided in the embodiment of the present application.

[0045] The following is an introduction to the advertising inventory inquiry method provided by this application through an embodiment.

[0046] See also Figure 3 , Figure 3 The following is a flowchart of the advertising inventory query method provided in the embodiment of the present application; for the convenience of description, the following embodiment is described with the server as the execution subject, it should be understood that the execution subject of the advertising inventory query method is not limited to the server, but can also be other devices with data processing functions such as terminal devices. Figure 3 As shown, the advertising inventory inquiry method includes the following steps:

[0047] Step 301: Determine the inventory corresponding to each ad slot on the target ad delivery page.

[0048] The target advertisement delivery page refers to the page used to deliver the target advertisement, and the target advertisement delivery page can usually be determined based on the targeting conditions proposed by the advertiser. The inventory corresponding to the ad slot refers to the remaining inventory of a certain ad slot that can be allocated for delivering the target advertisement; usually, the inventory of the delivered advertisements can be allocated first to ensure that the delivery demand of the delivered advertisements is met; then, based on the remaining inventory after the ad slot is allocated, the inventory of the ad slot that can be allocated for delivering the target advertisement is determined as the inventory corresponding to the ad slot.

[0049] When the server needs to query the advertising inventory for the target advertisement, the server can accordingly determine the inventory corresponding to each advertising position on the target advertisement delivery interface; it should be understood that the inventory corresponding to each advertising position on the target advertisement delivery interface can be the same or different.

[0050] Step 302: Determine the theoretical available inventory corresponding to the target advertisement to be delivered.

[0051] After determining the inventory corresponding to each ad slot on the target advertisement delivery page, the server can further determine the theoretical available inventory corresponding to the target advertisement based on the inventory corresponding to each ad slot and the targeting conditions of the target advertisement; wherein the targeting conditions of the target advertisement are usually set by the advertiser, which can characterize the delivery conditions of the target advertisement, such as delivery time, delivery object, delivery location, etc.

[0052] It should be noted that the above theoretical available inventory is actually the available inventory determined for the target advertisement without considering the deduplication rules. The theoretical available inventory is usually too large, and directly using the theoretical available inventory as the advertising inventory query result corresponding to the target advertisement will usually lead to the phenomenon of missing the target advertisement.

[0053] In a possible implementation, the server may determine the theoretical available inventory corresponding to the target advertisement according to the delivery conditions of the target advertisement to be delivered by using a high water mark (HWM) algorithm.

[0054] When the server determines the theoretical available inventory corresponding to the target advertisement through the HWM algorithm, it can first determine the priority of the target advertisement. Specifically, the server can first calculate the traffic sum of each traffic node that meets the conditions for placing each contract advertisement, and then determine the priority of the contract advertisement according to the traffic sum corresponding to each contract advertisement. The smaller the traffic sum, the higher the corresponding priority. In this way, the priority of the target advertisement and each contract advertisement is determined.

[0055] Then, the HWM algorithm is used to determine the theoretical available inventory. At this time, it is usually necessary to select different inventory values ​​for testing one by one, that is, to take different available inventories and substitute them into the HWM algorithm, and finally find the maximum value that makes the HWM allocation algorithm have a solution, as the theoretical available inventory corresponding to the target advertisement. Specifically, firstly, the inventory is allocated to the contract advertisements with a higher priority than the target advertisement, and the remaining inventory is allocated to the target advertisement to obtain the maximum available inventory Max; then, the priority of the target advertisement is set to the lowest, and the inventory is allocated to other high-priority contract advertisements, and the remaining inventory is allocated to the target advertisement to obtain the minimum available inventory Min. Then, a binary search is performed between the maximum available inventory Max and the minimum available inventory Min, and the solution is iterated into the allocation algorithm to find the maximum value that makes the HWM allocation algorithm have a solution, which is the theoretical available inventory corresponding to the target advertisement.

[0056] The specific calculation process is as follows:

[0057] Assume that the initial value of theoretical available inventory is I(0) = (Max+Min) / 2;

[0058] If the HWM algorithm has a solution, then I(t+1)=(Max+I(t)) / 2;

[0059] If the HWM algorithm has no solution, then I(t+1)=(Min+I(t)) / 2;

[0060] The solution is iterated in this way until I(t) converges, and finally the HWM allocation algorithm has a maximum value that is the theoretical available inventory corresponding to the target advertisement.

[0061] It should be noted that, in the process of determining the theoretical available inventory using the HWM algorithm, the advertisement playing probability corresponding to the available inventory will be obtained each time the HWM algorithm is executed.

[0062] Advertisement playing probability P = I(t) / Sj;

[0063] Correspondingly, the advertisement playing probability corresponding to the target advertisement is the ratio of the theoretical available inventory corresponding to the target advertisement to the sum of its corresponding traffic.

[0064] It should be understood that in addition to using the HWM algorithm to determine the theoretical available inventory corresponding to the target advertisement, the theoretical available inventory corresponding to the target advertisement may also be determined by other methods, and no limitation is imposed on the method for determining the theoretical available inventory.

[0065] Step 303: The theoretical available inventory corresponding to the target advertisement is corrected by a depreciation coefficient to obtain the actual available inventory corresponding to the target advertisement. The depreciation coefficient represents the degree of influence of the advertisement delivery system on the advertisement exposure when the deduplication rule is used for advertisement delivery.

[0066] After the server determines the theoretical available inventory corresponding to the target advertisement, it can use the impairment coefficient to correct the theoretical available inventory corresponding to the target advertisement, thereby obtaining the actual available inventory corresponding to the target advertisement; the impairment coefficient can characterize the degree of influence of the advertisement delivery system on the advertisement exposure when the deduplication rules are used for advertisement delivery.

[0067] Specifically, assuming that the depreciation coefficient is Q, the theoretical available inventory corresponding to the target advertisement is R O , then the actual available inventory R corresponding to the target advertisement d Theoretical available inventory R O Multiply it by the loss factor Q to get R d =R o ×Q.

[0068] In one possible implementation, the server can obtain the impairment coefficient corresponding to the target advertisement according to the advertisement playback probability corresponding to the target advertisement through a pre-trained linear regression model; the pre-trained linear regression model is generated by training a training data set, and the training data set includes multiple training data, and the training data includes the advertisement playback probability and the impairment coefficient corresponding to the advertisement.

[0069] The inventors have found that there is a specific relationship between the loss coefficient and the probability of advertisement playback, that is, the loss coefficient is a polynomial of the probability of advertisement playback. Figure 4 The relationship between the impairment coefficient and the probability of ad playback is derived. It is assumed that the target ad delivery page includes 3 ad slots, the inventory corresponding to each ad slot is S, the deduplication window range corresponding to the target ad delivery page is 3, and the ad playback probability corresponding to the target ad A to be delivered is P.

[0070] Theoretical available inventory R corresponding to target ad A o =3×S×P;

[0071] The actual available inventory R corresponding to the target advertisement d Equal to the theoretical inventory R corresponding to the target advertisement O Subtract the loss of inventory R caused by the deduplication rule S That is, R d =R o -R s .

[0072] Among them, the loss inventory R caused by the deduplication rule S It is the sum of the inventory loss caused by the deduplication rule on ad slot 2 and the inventory loss caused by the deduplication rule on ad slot 3.

[0073] That is, R S = Inventory loss due to deduplication of ad slot 2 + inventory loss due to deduplication of ad slot 3

[0074] =S×(the probability of ad A appearing in ad slots 1 and 2)+S×(the probability of ad A appearing in ad slots 1 and 3)+S×(the probability of ad A appearing in ad slots 2 and 3)

[0075] =S×P×P+S×P×P+S×(1-P)×P×P

[0076] and then,

[0077] Right now,

[0078] Further extended to the case where the number of ad slots is greater than 3 and the inventory corresponding to each ad slot is different, the depreciation coefficient Q can be expressed as: Q = w0 + w1P + w2P2 +…+w n P n

[0079] Based on the relationship between the above-mentioned impairment coefficient and the probability of advertisement playing, a machine learning algorithm can be used to utilize training data including the probability of advertisement playing and its corresponding impairment coefficient to train a linear regression model that can determine the impairment coefficient based on the probability of advertisement playing. The specific method of training the linear regression model will be introduced in the subsequent method embodiments. Please refer to the subsequent embodiments for details.

[0080] Furthermore, when the server needs to use the depreciation coefficient to correct the theoretical available inventory corresponding to the target advertisement, the server can input the advertisement playing probability corresponding to the target advertisement into the above-mentioned linear regression model, and use the linear regression model to process the advertisement playing probability corresponding to the target advertisement accordingly to obtain the correction coefficient corresponding to the advertisement playing probability, and then use the correction coefficient to correct the theoretical available inventory corresponding to the target advertisement.

[0081] In another possible implementation, the server may also collect historical delivery data of the advertising delivery system, where the historical delivery data includes actual exposure of the advertisements and scheduled exposure of the advertisements during the historical delivery time period; and then, determine the above-mentioned impairment coefficient based on the ratio of the actual exposure of the advertisements to the scheduled exposure of the advertisements in the historical delivery data.

[0082] Specifically, the server can use a certain time period in the past as a historical delivery time period according to actual needs, for example, use the previous month as the historical delivery time period; then, collect historical delivery data within the historical delivery time period in the advertising delivery system, and the historical delivery data generally includes actual exposures corresponding to multiple advertisements, and scheduled exposures corresponding to each of these advertisements; then, for a certain advertisement, divide its actual exposure by its scheduled exposure to obtain the loss coefficient corresponding to the advertisement, and in this way, determine the loss coefficient corresponding to each advertisement, and finally, determine the loss coefficient corresponding to the advertising delivery system based on the loss coefficient corresponding to each advertisement.

[0083] It should be noted that, in actual applications, the historical delivery data may only include the actual exposure and the scheduled exposure corresponding to an advertisement, and no limitation is imposed on the amount of data included in the historical delivery data.

[0084] It should be understood that in addition to the above two methods of determining the depreciation coefficient, in practical applications, other methods can be used to determine the depreciation coefficient according to actual needs, and no limitation is made here on the method of determining the depreciation coefficient.

[0085] When determining the advertising inventory inquiry results, the above-mentioned advertising inventory inquiry method uses the impairment coefficient to correct the theoretical available inventory corresponding to the advertisement. The impairment coefficient can characterize the degree of influence of the advertising delivery system using deduplication rules for advertising delivery on the advertising exposure. In this way, the influence of deduplication rules on the inquiry results is considered in the process of determining the advertising inventory inquiry results, thereby ensuring that the final inquiry results are more accurate, reducing the occurrence of advertising shortages, and ensuring that the actual needs of advertisers are met.

[0086] exist Figure 3 In the embodiment shown, it is mentioned that the server can modify the theoretical available inventory corresponding to the target advertisement by using a depreciation coefficient. In one possible implementation, the depreciation coefficient can be determined by using a linear regression model. To facilitate understanding of the determination method of the depreciation coefficient, the following is combined with Figure 5 This paper introduces the training method of linear regression model.

[0087] See also Figure 5 , Figure 5 A flowchart of the linear regression model training method provided in the embodiment of the present application; for ease of description, the following embodiment is described with the server as the execution subject, it should be understood that the execution subject of the linear regression model training method is not limited to the server, but can also be other devices with model training functions such as terminal devices; it should be noted that the server used to train the linear regression model and the server used to inquire about the advertising inventory can be the same server or different servers. Figure 5 As shown, the training method of the linear regression model includes the following steps:

[0088] Step 501: Determine a training data set.

[0089] When the server trains a linear regression model, it usually needs to use multiple training data to train the linear regression model; therefore, before training the linear regression model, the server needs to first obtain multiple training data to form a training data set. It should be understood that since the linear regression model is a model for determining the loss coefficient based on the probability of advertisement playback, the training data obtained by the server includes the probability of advertisement playback and its corresponding loss coefficient.

[0090] In one possible implementation, the server may collect training data in the following manner to form a training data set: first, obtain the probability of ad playback when the ad is delivered, the actual exposure corresponding to the ad, and the loss of exposure caused by deduplication when the ad is delivered; then, determine the correction coefficient corresponding to the ad based on the actual exposure and loss of exposure corresponding to the ad; and then, use the ad playback probability corresponding to the ad and the determined correction coefficient as a set of training data, and form a training data set through multiple sets of different training data.

[0091] Specifically, the actual exposure corresponding to an advertisement refers to the exposure actually generated in the process of delivering the advertisement; the loss exposure corresponding to an advertisement refers to the exposure lost due to certain reasons why the advertisement was not exposed. When loss exposure occurs, the advertising delivery platform usually records the reasons for the loss exposure accordingly. When collecting training data, the server can filter out the loss exposure caused by deduplication rules from all loss exposures, and use the loss exposure and the actual exposure corresponding to the advertisement to calculate the depreciation coefficient.

[0092] When calculating the impairment coefficient using the lost exposures and the actual exposures corresponding to the advertisements, you can first calculate the sum of the lost exposures caused by the deduplication rules and the actual exposures, and then divide the actual exposures by the sum of the lost exposures caused by the deduplication rules and the actual exposures to get the impairment coefficient.

[0093] The advertisement playing probability corresponding to the advertisement and the impairment coefficient obtained by the above calculation are used as a set of training data; in this way, multiple sets of different training data are obtained according to the above method, and these training data are used to form a training data set.

[0094] It should be understood that in actual applications, the server may also obtain the training data set in other ways, and no limitation is imposed on the way of obtaining the training data set here.

[0095] Step 502: A linear regression model is obtained by training the training data set using the least squares method.

[0096] After the server determines the training data set, it can train the initial linear regression model based on the training data set using the least square method to obtain a linear regression model that can accurately determine the correction coefficient according to the advertisement playing probability.

[0097] It should be noted that the least squares method is a mathematical optimization technique that finds the best function match between data by minimizing the sum of squares of errors. The least squares method can be used to easily obtain unknown data and minimize the sum of squares of errors between the obtained data and the actual data.

[0098] Specifically, when the linear regression model is trained by the least squares method, the server can use each exponential item of the advertisement playing probability in the training data as a feature item, for example, the exponential items P, P^2, P^3, ..., P^n of the advertisement playing probability P as feature items; then, the linear regression model is trained by the least squares method, and the specific expression of the loss function used in the training process is shown in formula (1):

[0099]

[0100] Formula (1) is a convex function, which has a minimum value. The training process of the linear regression model is actually to find the coefficient vector W that makes Formula (1) reach the minimum value. During the specific training, the gradient descent method can be used to solve it. A set of random numbers are selected as the initial value of the coefficient vector W. The coefficient vector W is corrected according to Formula (2) and Formula (3) using the training data:

[0101]

[0102]

[0103] Wherein, a is the learning rate, which can be selected as a smaller value according to the actual situation; the above formulas (2) and (3) are iteratively executed until the coefficient W converges, thus obtaining the coefficient vector W.

[0104] The initial linear regression model is trained with different parameters, and then the root mean square errors of different linear regression models are compared. Usually, when the polynomial order n = 2 and the regularization coefficient λ = 0.0001, the calculated model root mean square error is the smallest. At this time, the linear regression model training can be considered complete.

[0105] It should be understood that in addition to the above-mentioned method, other methods can also be used to determine whether the linear regression model is trained. No limitation is made here on the method of determining whether the linear regression model is trained.

[0106] It should be noted that when the deduplication window changes, the previously trained linear regression model will no longer be applicable to the advertising delivery page after the deduplication window changes. Therefore, after the deduplication window changes, it is necessary to retrain the linear regression model applicable to the new deduplication window; at this time, the server needs to collect another training data set, retrain the linear regression model based on the re-collected training data set, and use the retrained linear regression model in the advertising delivery system to determine the loss coefficient of the advertisement to be delivered.

[0107] The training method used when retraining the linear regression model is the same as above Figure 5 The training method of the linear regression model shown is similar, and it is necessary to collect the actual exposure corresponding to the advertisement after the deduplication window changes and the loss exposure of the advertisement caused by the deduplication rules, and then calculate the impairment coefficient corresponding to the advertisement based on the actual exposure and the loss exposure in the above manner, and use the impairment coefficient and the advertisement playback probability corresponding to the advertisement as a set of training data. In this way, multiple sets of training data generated after the deduplication window changes are obtained to form a new training data set; then, based on the new training data set, the least squares method is used to train a new linear regression model.

[0108] The training method of the above linear regression model is based on a plurality of training data including advertisement play probability and its corresponding impairment coefficient, and the least square method is used to train a linear regression model, which can accurately determine the corresponding impairment coefficient according to the input advertisement play probability. The training data used to train the linear regression model are all determined based on the historical advertisement play data, which is relatively objective and accurate. Accordingly, the linear regression model trained based on these training data can also determine the impairment coefficient relatively accurately. Furthermore, when the impairment coefficient determined by the linear regression model is used to correct the theoretical available inventory, it can ensure that the actual available inventory is accurate and reasonable.

[0109] In order to facilitate further understanding of the advertising inventory inquiry method provided in the embodiment of the present application, the following takes the server using a linear regression model to determine the depreciation coefficient as an example to provide an overall introduction to the advertising inventory inquiry method provided in the embodiment of the present application.

[0110] If a linear regression model is to be used to determine the impairment coefficient, the server needs to pre-train a linear regression model that can be applied to the current deduplication window range; specifically, the server can first collect multiple groups of training data to form a training data set, and the collected training data include the probability of advertisement playback and its corresponding impairment coefficient, which can be calculated based on the actual exposure of the advertisement and the loss of exposure caused by the deduplication rules; then, the server can use the least squares method to train a linear regression model based on the above training data set, and the linear regression model can determine its corresponding impairment coefficient according to the input advertisement playback probability.

[0111] It should be noted that when the deduplication window range on the advertising delivery page changes, the server can re-collect multiple training data to form a new training data set after the deduplication window range changes, and then use the least squares method to retrain the linear regression model in combination with the training data set.

[0112] When an advertiser needs to deliver a target advertisement through an advertisement delivery platform, the server can first determine the inventory corresponding to each ad slot on the target advertisement delivery page, as well as the theoretical available inventory corresponding to the target advertisement. When specifically determining the theoretical available inventory corresponding to the target advertisement, the server can use the HWM algorithm to determine the theoretical available inventory corresponding to the target advertisement based on the delivery conditions, i.e., the targeting conditions, of the target advertisement. At the same time, the advertisement playback probability corresponding to the target advertisement can also be calculated.

[0113] It should be understood that the above theoretical available inventory is the available inventory calculated without considering the deduplication rules, which is usually too high. Directly using the theoretical available inventory as the query result can easily lead to the occurrence of advertisement gaps.

[0114] Furthermore, the server uses the linear regression model trained through the above training process to determine the corresponding impairment coefficient according to the advertisement playback probability corresponding to the target advertisement. Since the training data used in training the linear regression model includes the advertisement playback probability, as well as the impairment coefficient determined based on the actual advertisement exposure and the loss exposure caused by the deduplication rules, the impairment coefficient determined by the linear regression model can correspondingly characterize the degree of impact of the deduplication rules on the advertisement exposure when the advertisement is placed.

[0115] Finally, the server uses the theoretical available inventory corresponding to the target advertisement and multiplies it by the depreciation coefficient determined by the linear regression model to obtain the actual available inventory corresponding to the target advertisement.

[0116] The inventors have experimentally found that when the advertising inventory inquiry method provided in the embodiment of the present application is applied to a news APP, the absence rate of its information flow advertisements is significantly reduced from 27% to 7%. It can be seen that the advertising inventory inquiry method can ensure that the determined inquiry results are relatively accurate, effectively reduce the absence rate of advertisements, and improve the advertiser's experience.

[0117] With respect to the advertising inventory inquiry method described above, the present application also provides a corresponding advertising inventory inquiry device, so that the above-mentioned advertising inventory inquiry method can be applied and implemented in practice.

[0118] See also Figure 6 , Figure 6 It is the same as above Figure 3 The structure diagram of an advertising inventory inquiry device 600 corresponding to the advertising inventory inquiry method shown in the figure, the advertising inventory inquiry device 600 includes:

[0119] The first determination module 601 is used to determine the inventory corresponding to each advertising position on the target advertising delivery page;

[0120] The second determination module 602 is used to determine the theoretical available inventory corresponding to the target advertisement to be delivered;

[0121] The correction module 603 is used to correct the theoretical available inventory corresponding to the target advertisement by using a depreciation coefficient to obtain the actual available inventory corresponding to the target advertisement. The depreciation coefficient represents the degree of influence of the advertisement delivery system on the advertisement exposure when the deduplication rule is used for advertisement delivery.

[0122] Optional, in Figure 6 Based on the advertising inventory inquiry device shown, see Figure 7 , Figure 7 Another advertising inventory inquiry device provided in an embodiment of the present application; the advertising inventory inquiry device 700 also includes:

[0123] The acquisition module 701 is used to obtain the impairment coefficient corresponding to the target advertisement according to the advertisement playing probability corresponding to the target advertisement through a pre-trained linear regression model, wherein the linear regression model is generated by training a training data set, and the training data set includes multiple training data, and the training data includes the advertisement playing probability and the impairment coefficient corresponding to the advertisement.

[0124] Optional, in Figure 7 Based on the advertising inventory inquiry device shown, see Figure 8 , Figure 8 Another advertising inventory inquiry device provided in an embodiment of the present application; the advertising inventory inquiry device 800 also includes:

[0125] A training data determination module 801 is used to determine the training data set;

[0126] The training module 802 is used to obtain the linear regression model by training the training data set through the least square method.

[0127] Optional, in Figure 8 Based on the advertising inventory inquiry device shown, the training data determination module 801 is specifically used for:

[0128] Get the probability of advertisement playing and the actual exposure of the advertisement when it is delivered;

[0129] Get the loss of impressions caused by deduplication during ad delivery;

[0130] Determine the loss coefficient corresponding to the advertisement based on the actual exposure and loss exposure corresponding to the advertisement;

[0131] The advertisement playing probability corresponding to the advertisement and the impairment coefficient corresponding to the advertisement are taken as a set of training data, and the training data set is formed by multiple sets of different training data.

[0132] Optional, in Figure 8 Based on the advertising inventory inquiry device shown, see Fig. 9 , Fig. 9 Another advertising inventory inquiry device provided in an embodiment of the present application; the advertising inventory inquiry device 900 also includes:

[0133] The updating training module 901 is used to collect another training data set based on the updated deduplication window when the deduplication window used by the advertising delivery system changes, and retrain the linear regression model based on the other training data set for use in the advertising delivery system.

[0134] Optional, in Figure 6 Based on the advertising inventory inquiry device shown, see Fig.10 , Fig.10 Another advertising inventory inquiry device provided in an embodiment of the present application; the advertising inventory inquiry device 1000 also includes:

[0135] The collection module 1001 is used to collect historical delivery data of the advertisement delivery system, wherein the historical delivery data includes the actual exposure amount of the advertisement and the scheduled exposure amount of the advertisement in the historical delivery time period;

[0136] The third determination module 1002 is used to determine the impairment coefficient according to the ratio of the actual exposure of the advertisement in the historical delivery data to the predetermined exposure of the advertisement.

[0137] Optional, in Figure 6 Based on the advertising inventory inquiry device shown, the second determination module is specifically used for:

[0138] The theoretical available inventory corresponding to the target advertisement to be delivered is determined through a high water mark algorithm according to delivery conditions of the target advertisement to be delivered.

[0139] When determining the advertising inventory inquiry result, the above-mentioned advertising inventory inquiry device uses the depreciation coefficient to correct the theoretical available inventory corresponding to the advertisement. The depreciation coefficient can characterize the degree of influence of the advertising delivery system on the advertising exposure when using deduplication rules for advertising delivery. In this way, in the process of determining the advertising inventory inquiry result, the influence of deduplication rules on the inquiry result is considered to ensure that the final inquiry result is more accurate, reduce the occurrence of advertising shortages, and ensure that the actual needs of advertisers are met.

[0140] The present application also provides a device for querying advertising inventory, which may specifically be a server, see Fig.11 , Fig.11 A schematic diagram of the structure of a server for querying advertising inventory provided in an embodiment of the present application. The server 1100 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPU) 1122 (for example, one or more processors) and memory 1132, and one or more storage media 1130 (for example, one or more mass storage devices) storing application programs 1142 or data 1144. Among them, the memory 1132 and the storage medium 1130 can be short-term storage or permanent storage. The program stored in the storage medium 1130 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server. Furthermore, the central processor 1122 can be configured to communicate with the storage medium 1130 to execute a series of instruction operations in the storage medium 1130 on the server 1100.

[0141] The server 1100 may also include one or more power supplies 1126, one or more wired or wireless network interfaces 1150, one or more input and output interfaces 1158, and / or one or more operating systems 1141, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0142] The steps performed by the server in the above embodiment can be based on the Fig.11 The server structure shown.

[0143] The CPU 1122 is used to execute the following steps:

[0144] Determine the inventory corresponding to each ad slot on the target ad delivery page;

[0145] Determine the theoretical available inventory corresponding to the target advertisement to be delivered;

[0146] The theoretical available inventory corresponding to the target advertisement is corrected by a depreciation coefficient to obtain the actual available inventory corresponding to the target advertisement. The depreciation coefficient represents the degree of influence on the advertisement exposure when the advertisement delivery system adopts deduplication rules for advertisement delivery.

[0147] Optionally, CPU 1122 may also execute method steps of any specific implementation of the advertising inventory inquiry method in the embodiments of the present application.

[0148] The present application also provides another device for querying advertising inventory, which may be a terminal device. Fig.12 For the sake of convenience, only the parts related to the embodiments of the present application are shown. For specific technical details not disclosed, please refer to the method part of the embodiments of the present application. The terminal can be any terminal device including a mobile phone, a tablet computer, a personal digital assistant (PDA), a point of sales (POS), a car computer, etc., taking the mobile phone as an example:

[0149] Fig.12 FIG. 1 is a block diagram showing a partial structure of a mobile phone related to a terminal provided in an embodiment of the present application. Fig.12The mobile phone includes: a radio frequency (RF) circuit 1210, a memory 1220, an input unit 1230, a display unit 1240, a sensor 1250, an audio circuit 1260, a wireless fidelity (WiFi) module 1270, a processor 1280, and a power supply 1290 and other components; wherein the input unit 1230 includes a touch panel 1231 and other input devices 1232, the display unit 1240 includes a display panel 1241, and the audio circuit 1260 includes a speaker 1261 and a microphone 1262.

[0150] Those skilled in the art will understand that Fig.12 The mobile phone structure shown in the figure does not constitute a limitation on the mobile phone, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0151] In the embodiment of the present application, the processor 1280 included in the terminal has the following functions:

[0152] Determine the inventory corresponding to each ad slot on the target ad delivery page;

[0153] Determine the theoretical available inventory corresponding to the target advertisement to be delivered;

[0154] The theoretical available inventory corresponding to the target advertisement is corrected by a depreciation coefficient to obtain the actual available inventory corresponding to the target advertisement. The depreciation coefficient represents the degree of influence on the advertisement exposure when the advertisement delivery system adopts deduplication rules for advertisement delivery.

[0155] Optionally, processor 1280 may also execute method steps of any specific implementation of the advertising inventory inquiry method in the embodiments of the present application.

[0156] The embodiment of the present application also provides a computer-readable storage medium for storing program code, wherein the program code is used to execute any one of the implementation methods of the advertising inventory inquiry method described in the aforementioned embodiments.

[0157] The embodiment of the present application also provides a computer program product including instructions, which, when executed on a computer, enables the computer to execute any one of the implementation methods of the advertising inventory inquiry method described in the aforementioned embodiments.

[0158] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0159] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0160] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0161] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0162] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (full name in English: Read-Only Memory, English abbreviation: ROM), random access memory (full name in English: Random Access Memory, English abbreviation: RAM), disk or optical disk and other media that can store program codes.

[0163] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for querying advertising inventory, characterized in that: include: Determine the inventory corresponding to each ad slot on the target ad delivery page, wherein the inventory corresponding to the ad slot refers to the remaining inventory of the ad slot allocated for delivering the target ad; Determine the theoretical available inventory corresponding to the target advertisement to be delivered, including: calculating the traffic sum of each traffic node that meets the delivery conditions of each contract advertisement; determining the priority of the contract advertisement according to the traffic sum corresponding to each contract advertisement, so as to determine the priority of the target advertisement and each contract advertisement; allocating inventory for the contract advertisement with a higher priority than the target advertisement, allocating the remaining inventory to the target advertisement, and obtaining the maximum available inventory; setting the priority of the target advertisement to the lowest, allocating inventory for other high-priority contract advertisements, allocating the remaining inventory to the target advertisement, and obtaining the minimum available inventory; performing a binary search between the maximum available inventory and the minimum available inventory, substituting it into the allocation algorithm for iterative solution to obtain the maximum value of the high water mark algorithm, and the maximum value is the theoretical available inventory corresponding to the target advertisement; The theoretical available inventory corresponding to the target advertisement is corrected by a depreciation coefficient to obtain the actual available inventory corresponding to the target advertisement, wherein the depreciation coefficient represents the degree of influence on the advertisement exposure when the advertisement delivery system adopts a deduplication rule for advertisement delivery, and the deduplication rule is that the same advertisement cannot be delivered within a certain window range; The depreciation coefficient is determined in the following way: Obtaining the impairment coefficient corresponding to the target advertisement according to the advertisement playing probability corresponding to the target advertisement through a pre-trained linear regression model, wherein the linear regression model is generated by training a training data set, the training data set includes a plurality of training data, the training data includes the advertisement playing probability and the impairment coefficient corresponding to the advertisement, and the advertisement playing probability is the ratio of the theoretical available inventory corresponding to the target advertisement to the sum of its corresponding traffic; or, Collecting historical delivery data of the advertisement delivery system, wherein the historical delivery data includes actual exposure of the advertisement and scheduled exposure of the advertisement within the historical delivery time period; The impairment coefficient is determined according to a ratio of the actual exposure of the advertisement in the historical delivery data to the predetermined exposure of the advertisement.

2. The method according to claim 1, characterized in that: The linear regression model is trained by: Determining the training data set; The linear regression model is obtained by training the training data set using the least squares method.

3. The method according to claim 1 or 2, characterized in that: The training data set is determined by the following method, including: Get the probability of advertisement playing and the actual exposure of the advertisement when it is delivered; Get the loss of impressions caused by deduplication during ad delivery; Determine the loss coefficient corresponding to the advertisement based on the actual exposure and loss exposure corresponding to the advertisement; The advertisement playing probability corresponding to the advertisement and the impairment coefficient corresponding to the advertisement are taken as a set of training data, and the training data set is formed by multiple sets of different training data.

4. The method according to claim 1, characterized in that: When the deduplication window used by the advertisement delivery system changes, another training data set is collected based on the updated deduplication window, and the linear regression model is retrained based on the another training data set for use in the advertisement delivery system.

5. An advertising inventory inquiry device, characterized in that: include: A first determination module is used to determine the inventory corresponding to each ad slot on the target ad delivery page, wherein the inventory corresponding to the ad slot refers to the remaining inventory of the ad slot allocated for delivering the target ad; The second determination module is used to determine the theoretical available inventory corresponding to the target advertisement to be delivered, which includes: calculating the traffic sum of each traffic node that meets the delivery conditions of each contract advertisement; determining the priority of the contract advertisement according to the traffic sum corresponding to each contract advertisement, so as to determine the priority of the target advertisement and each contract advertisement; allocating inventory for the contract advertisement with a higher priority than the target advertisement, and allocating the remaining inventory to the target advertisement to obtain the maximum available inventory; setting the priority of the target advertisement to the lowest, allocating inventory for other high-priority contract advertisements, and allocating the remaining inventory to the target advertisement to obtain the minimum available inventory; performing a binary search between the maximum available inventory and the minimum available inventory, substituting it into the allocation algorithm to iteratively solve the maximum value of the high water mark algorithm, and the maximum value is the theoretical available inventory corresponding to the target advertisement; A correction module, used to correct the theoretical available inventory corresponding to the target advertisement by using a depreciation coefficient to obtain the actual available inventory corresponding to the target advertisement, wherein the depreciation coefficient represents the degree of influence on the advertisement exposure when the advertisement delivery system adopts a deduplication rule for advertisement delivery, and the deduplication rule is that the same advertisement cannot be delivered within a certain window range; The device also includes: An acquisition module is used to obtain the impairment coefficient corresponding to the target advertisement according to the advertisement playing probability corresponding to the target advertisement through a pre-trained linear regression model, wherein the linear regression model is generated by training a training data set, the training data set includes a plurality of training data, the training data includes the advertisement playing probability and the impairment coefficient corresponding to the advertisement, and the advertisement playing probability is the ratio of the theoretical available inventory corresponding to the target advertisement to the sum of its corresponding traffic; or, A collection module, used to collect historical delivery data of the advertisement delivery system, wherein the historical delivery data includes actual exposure of advertisements and scheduled exposure of advertisements within a historical delivery time period; The fourth determination module is used to determine the impairment coefficient according to the ratio of the actual exposure of the advertisement in the historical delivery data to the predetermined exposure of the advertisement.

6. The device according to claim 5, characterized in that The device also includes: A training data determination module, used to determine the training data set; The training module is used to obtain the linear regression model by training the training data set through the least square method.

7. The device according to claim 5 or 6, characterized in that The training data determination module is specifically used for: Get the probability of advertisement playing and the actual exposure of the advertisement when it is delivered; Get the loss of impressions caused by deduplication during ad delivery; Determine the loss coefficient corresponding to the advertisement based on the actual exposure and loss exposure corresponding to the advertisement; The advertisement playing probability corresponding to the advertisement and the impairment coefficient corresponding to the advertisement are taken as a set of training data, and the training data set is formed by multiple sets of different training data.

8. The device according to claim 5, characterized in that The device also includes an update training module; The update training module is used to collect another training data set based on the updated deduplication window when the deduplication window used by the advertising delivery system changes, and retrain the linear regression model based on the other training data set for use in the advertising delivery system.

9. A device, characterized in that: The device comprises a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the advertising inventory inquiry method according to any one of claims 1-4 according to the instructions in the program code.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium is used to store program code, and the program code is used to execute the advertising inventory query according to any one of claims 1-4.

11. A computer program product comprising instructions, characterized in that When it is run on a computer, the computer is enabled to execute the advertising inventory inquiry method according to any one of claims 1 to 4.

Images