Information recommendation method and device, electronic equipment, computer readable storage medium and computer program product

Abstract

The invention provides an information recommendation method and device, electronic equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining a recommendation information set and a flow set used for releasing the recommendation information set, and obtaining original cost for releasing the recommendation information set to the flow set and a start-up budget of the recommendation information set competing for the flow set through a one-key start-up strategy; in response to the received instruction of starting the one-key quantity starting strategy, determining a quantity starting factor corresponding to each piece of recommendation information in the recommendation information set based on the recommendation information set, the flow set and the quantity starting budget; controlling the recommendation information set to compete for the flow set based on the original cost and the start-up factor to obtain a competition result; and based on the competition result, putting the recommendation information set into the traffic set to obtain a maximum increment conversion total amount. According to the method and the device, the total increment conversion amount of the recommendation information can be maximized within the range of the initial budget.

Description

Technical Field

[0001] This application relates to computer technology, and more particularly to an information recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product. Background Technology

[0002] In advertising scenarios, advertisers often need to quickly test new content and accelerate ad launches, meaning they want to quickly verify the quality of their creative materials, increase bids to boost traffic in a short period, and help ads quickly overcome the cold start phase. One-click scaling is a product designed to meet these needs. Advertisers allocate a separate budget for scaling their ads and spend it within a certain timeframe to quickly acquire users, accumulate data for their models, and quickly overcome the cold start period. Therefore, the key to the one-click scaling strategy is to spend the advertiser's allocated scaling budget within a limited timeframe at the lowest possible cost while maximizing total conversions. Summary of the Invention

[0003] This application provides an information recommendation method, apparatus, electronic device, computer-readable storage medium, and computer program product, which can maximize the total incremental conversion of recommended information within a budget.

[0004] The technical solution of this application embodiment is implemented as follows:

[0005] This application provides an information recommendation method, the method comprising:

[0006] Obtain a set of recommended information, a set of traffic for delivering the set of recommended information, and obtain the original cost of delivering the set of recommended information to the set of traffic, as well as the budget for the set of recommended information to compete for the set of traffic through a one-click scaling strategy.

[0007] In response to receiving an instruction to enable a one-click scaling strategy, based on the recommended information set, the traffic set, and the scaling budget, a scaling factor is determined for each recommended information in the recommended information set. The scaling factor is used to determine the scaling cost when the recommended information competes for the traffic set.

[0008] Based on the original cost and the scaling factor, the recommended information set competes with the traffic set to obtain the competition result;

[0009] Based on the competition results, the set of recommended information is distributed to the set of traffic to obtain the maximum total incremental conversion.

[0010] This application provides an information recommendation device, the device comprising:

[0011] The acquisition module is used to acquire a set of recommendation information, a set of traffic for delivering the set of recommendation information, the original cost of delivering the set of recommendation information to the set of traffic, and the initial budget for the set of recommendation information to bid for the set of traffic through a one-click scaling strategy.

[0012] The determination module is used to respond to receiving an instruction to enable the one-click scaling strategy, and to determine the scaling factor corresponding to each piece of recommendation in the recommendation information set based on the recommendation information set, the traffic set, and the scaling budget. The scaling factor is used to determine the scaling cost when the recommendation information bids for the traffic set.

[0013] The control module is used to control the bidding of the recommended information set for the traffic set based on the original cost and the volume factor, so as to obtain the competition result;

[0014] The delivery module is used to deliver the set of recommended information to the set of traffic based on the competition results, so as to obtain the maximum incremental conversion amount.

[0015] This application provides an electronic device, the electronic device comprising:

[0016] Memory is used to store executable instructions for a computer;

[0017] The processor, when executing computer-executable instructions stored in the memory, implements the information recommendation method provided in the embodiments of this application.

[0018] This application provides a computer-readable storage medium storing a computer program or computer-executable instructions for implementing the information recommendation method provided in this application when executed by a processor.

[0019] This application provides a computer program product, including a computer program or computer executable instructions, which, when executed by a processor, implements the information recommendation method provided in this application.

[0020] The embodiments of this application have the following beneficial effects:

[0021] The system acquires a set of recommended information, a set of traffic for delivering the recommended information, the initial cost of delivering the recommended information to the traffic set, and the initial budget for bidding on the traffic set using a one-click scaling strategy. Upon receiving an instruction to enable the one-click scaling strategy, it determines the scaling factor for each recommended information within the set based on the recommended information, the traffic set, and the scaling budget. When bidding on the traffic set for the recommended information, it controls the cost of bidding for traffic for each recommended information based on the initial cost and the scaling factor. When the initial cost of a recommended information is too low to attract traffic, the cost is increased to the initial cost using the scaling factor, thereby enhancing the competitiveness of the recommended information and helping it acquire traffic. Simultaneously, the scaling factor is controlled to avoid excessively high initial cost increases that would waste the scaling budget, thus achieving the goal of acquiring more traffic at the lowest possible cost within the scaling budget. This maximizes the incremental conversion amount after the recommended information is delivered to the traffic set, thereby improving the efficiency and profitability of information recommendations. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the architecture of the information recommendation system 100 provided in an embodiment of this application;

[0023] Figure 2 This is a schematic diagram of the structure of the server 200 provided in the embodiments of this application;

[0024] Figure 3A This is a flowchart illustrating the information recommendation method provided in an embodiment of this application;

[0025] Figure 3B This is a flowchart illustrating the method for determining the scaling factor provided in an embodiment of this application;

[0026] Figure 3C This is a flowchart illustrating the contention traffic aggregation method provided in an embodiment of this application;

[0027] Figure 3D This is a flowchart illustrating the traffic differentiation method provided in an embodiment of this application;

[0028] Figure 3E This is a flowchart illustrating the updated budget method provided in an embodiment of this application;

[0029] Figure 4 This is a schematic diagram illustrating the effect of the one-click volume increase strategy provided in the embodiments of this application;

[0030] Figure 5A This is a schematic diagram of the non-one-click volume increase strategy provided in the embodiments of this application;

[0031] Figure 5BThis is a schematic diagram of the one-click scaling strategy (one) provided in the embodiments of this application;

[0032] Figure 5C This is a schematic diagram of the one-click scaling strategy (II) provided in the embodiments of this application. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0034] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0035] In the following description, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0036] In this application embodiment, the terms "module" or "unit" refer to a computer program or part of a computer program that has a predetermined function and works with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0037] Unless otherwise defined, all technical and scientific terms used in the embodiments of this application have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in the embodiments of this application is for the purpose of describing the embodiments of this application only and is not intended to limit this application.

[0038] In the implementation of this application, the collection and processing of relevant data should strictly comply with the requirements of relevant national laws and regulations, obtain the informed consent or separate consent of the personal information subject, and carry out subsequent data use and processing within the scope of laws and regulations and the authorization of the personal information subject.

[0039] Before providing a further detailed description of the embodiments of this application, the nouns and terms involved in the embodiments of this application will be explained, and the nouns and terms involved in the embodiments of this application shall be interpreted as follows.

[0040] 1) One-Click Scaling Strategy: This is a scaling product where the client allocates a separate scaling budget to the ad campaign. The system aims to spend this budget within a certain timeframe, helping the client quickly acquire traffic, accumulate data for the model, and quickly overcome the cold start period. An agreement is reached with the client that during the one-click scaling period, scaling costs are not guaranteed, and there is no compensation policy. The one-click scaling strategy is only used to acquire incremental traffic—traffic that cannot be obtained with the original cost—and the cost is deducted from the one-click scaling budget. Traffic that can be acquired with the original cost is no longer targeted by the one-click scaling strategy and is deducted from the original budget. This model is called the incremental mode (or Delta mode).

[0041] 2) Massive Advertising: This is an advertising strategy typically used by e-commerce platforms or advertisers who want to rapidly increase ad exposure by purchasing large quantities of goods or services. Through mass advertising, advertisers can quickly increase exposure within a certain period by rapidly increasing spending or ad frequency, thereby increasing sales or brand awareness. This method may require a large initial investment, but can yield significant returns once successfully scaled up.

[0042] 3) Incremental traffic: Traffic that can only be obtained through cost competition after the one-click traffic scaling strategy is activated and within the effective period of the one-click traffic scaling strategy.

[0043] 4) Existing traffic: After the one-click traffic boosting strategy is activated, within the effective period of the one-click traffic boosting strategy, after removing the one-click traffic boosting factor, the traffic can be obtained by bidding according to the original cost without using the one-click traffic boosting strategy.

[0044] 5) Non-one-click traffic: For ads that have enabled the one-click traffic strategy, traffic obtained through competition at the original cost outside of the 6 hours when the one-click traffic is enabled.

[0045] 6) Initial budget: The total amount of advertising expenditure set in advance for non-one-click scaling strategies before advertising is launched.

[0046] 7) Initial budget: The total amount of advertising expenditure that is pre-set before launching an advertising campaign to implement the one-click initial budget strategy.

[0047] 8) Original cost: The expected cost of competing for traffic for the advertisement, set in advance before launching the advertisement, without using the one-click scaling strategy.

[0048] 9) Scaling Cost: The actual cost of advertising traffic when using the one-click scaling strategy during the advertising competition traffic period. The scaling cost is determined based on the scaling factor and the original cost.

[0049] 10) Start-up Factor: A value set to increase the competitiveness of advertising when competing for traffic. By adjusting the start-up factor, the cost of starting an advertising campaign in the process of competing for traffic can be adjusted.

[0050] 11) Gross Merchandise Volume (GMV): The total value of all transactions within a certain time frame, including the sales of all goods or services, excluding refunds, discounts, or order cancellations. It is an important indicator for measuring the overall transaction value of a trading platform.

[0051] 12) Effective Cost Per Mille (ECPM): The average cost an advertiser needs to pay to obtain one thousand ad impressions.

[0052] One-click scaling strategies aggressively bid within a target timeframe to help recommended content gain more traffic. However, since scaling costs are not guaranteed during the one-click scaling period, they can lead to excessively high costs and wasted scaling budgets. Conversely, excessively low scaling costs may result in the recommended content failing to gain any traffic. Therefore, the key challenge of one-click scaling strategies is how to spend the advertiser's (the target audience of the recommended content) scaling budget with the lowest possible scaling cost and gain the most traffic within a limited timeframe.

[0053] This application provides an information recommendation method, apparatus, device, computer-readable storage medium, and computer program product. When recommended information cannot acquire traffic through bidding at its original cost, a scaling factor is used to increase the bidding cost to the scaling cost, thereby enhancing the competitiveness of the recommended information and helping it acquire traffic. Simultaneously, it avoids excessively increasing the scaling cost through the scaling factor, thus avoiding wasting the scaling budget. This achieves the goal of acquiring more traffic at the lowest possible scaling cost within the scaling budget, maximizing the incremental conversion amount after the recommended information set is deployed to the traffic set, thereby improving the efficiency and profitability of information recommendation. The following describes exemplary applications of the electronic devices provided in this application. These devices can be implemented as various types of terminals such as laptops, tablets, desktop computers, set-top boxes, smartphones, smart speakers, smartwatches, smart TVs, and in-vehicle terminals, or as servers. The following describes exemplary applications when the electronic device is implemented as a server.

[0054] See Figure 1 , Figure 1This is a schematic diagram of the architecture of the information recommendation system 100 provided in the embodiments of this application. The terminal 400 is connected to the server 200 through the network 300. The terminal 400 includes a graphical interface 410. The network 300 can be a wide area network or a local area network, or a combination of the two.

[0055] Server 200 is used to acquire a set of recommended information, a set of traffic for delivering the recommended information, the original cost of delivering the recommended information to the traffic set, and the scaling budget for the recommended information to compete for traffic through a one-click scaling strategy. The recipient of the recommended information executes the operation of enabling the one-click scaling strategy through the graphical interface 410 of terminal 400, and then the terminal 400 generates an instruction to enable the one-click scaling strategy and sends it to server 200. Upon receiving the instruction to enable the one-click scaling strategy, server 200 determines the scaling factor corresponding to each recommended information in the recommended information set based on the recommended information set, the traffic set, and the scaling budget. Then, based on the original cost and the scaling factor, it controls the competition between the recommended information set and the traffic set, obtains the competition results, and delivers the recommended information to the traffic set based on the competition results. After the recommended information is delivered to the traffic set, the recommended information delivered to the traffic set is displayed through the graphical interface 410 of terminal 400. When the recipient of the recommended information performs a specific action on the recommended information (such as purchasing a product, clicking a link, downloading an application, etc.) in the graphical interface 410, an incremental conversion total is generated.

[0056] In some embodiments, server 200 may be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. Terminals and servers can be connected directly or indirectly via wired or wireless communication, which is not limited in this embodiment.

[0057] See Figure 2 , Figure 2 This is a schematic diagram of the structure of the server 200 provided in the embodiments of this application. Figure 2 The server 200 shown includes at least one processor 210, memory 230, and at least one network interface 220. The various components of server 200 are coupled together via a bus system 240. It is understood that the bus system 240 is used to implement communication between these components. In addition to a data bus, the bus system 240 also includes a power bus, a control bus, and a status signal bus. However, for clarity, ... Figure 2The general labeled all buses as Bus System 240.

[0058] The processor 210 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0059] The memory 230 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drives, optical disk drives, etc. The memory 230 may optionally include one or more storage devices physically located away from the processor 210.

[0060] The memory 230 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), and the volatile memory may be random access memory (RAM). The memory 230 described in this application embodiment is intended to include any suitable type of memory.

[0061] In some embodiments, memory 230 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, as illustrated below.

[0062] Operating system 231 includes system programs for handling various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., for implementing various basic business functions and handling hardware-based tasks;

[0063] The network communication module 232 is used to reach other electronic devices via one or more (wired or wireless) network interfaces 220, exemplary network interfaces 220 including Bluetooth, WiFi, and Universal Serial Bus (USB).

[0064] In some embodiments, the apparatus provided in this application can be implemented in software. Figure 2An information recommendation device 233 stored in memory 230 is shown. This device can be software in the form of programs and plug-ins, and includes the following software modules: an acquisition module 2331, a determination module 2332, a control module 2333, and a delivery module 2334. These modules are logically connected and can therefore be arbitrarily combined or further separated according to their implemented functions. The functions of each module will be described below.

[0065] In other embodiments, the apparatus provided in this application can be implemented in hardware. As an example, the apparatus provided in this application can be a processor in the form of a hardware decoding processor, which is programmed to execute the information recommendation method provided in this application. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0066] The information recommendation method provided in this application will be described in conjunction with exemplary applications and implementations of the server provided in the embodiments of this application.

[0067] The information recommendation method provided in the embodiments of this application will be described below. As mentioned above, the electronic device implementing the information recommendation method in the embodiments of this application is a server. Therefore, the executing entity of each step will not be described again below.

[0068] It should be noted that the examples of information recommendation methods in the following text are illustrated using advertising information as an example. Those skilled in the art can apply the information recommendation methods provided in the embodiments of this application to the processing of other types of information recommendations based on their understanding of the following text.

[0069] See Figure 3A , Figure 3A This is a flowchart illustrating the information recommendation method provided in the embodiments of this application, which will be combined with... Figure 3A The steps shown are explained.

[0070] In step 101, the recommended information set, the traffic set used to deliver the recommended information set, the original cost of delivering the recommended information set to the traffic set, and the scaling budget for the recommended information set to compete for the traffic set through the one-click scaling strategy are obtained.

[0071] Here, the recommended ad set, traffic set, initial cost, and initial budget are pre-set. The recommended ad set is the collection of ads to be placed, which can be ads on various themes such as apps, movies, music, food, fashion trends, home décor, and sports, used to convey product or service information to the public, helping consumers understand and recognize the brand and products, and stimulating consumer purchases, downloads, and other behaviors. By placing recommended ads, corresponding conversions (conversions generated by consumer purchases, downloads, etc.) can be obtained, attracting potential viewers of the recommended ads to generate actual conversions, thereby improving sales performance and brand influence. The traffic set is the traffic that recommended ads can compete for through bidding for digital advertising, such as search engine traffic (traffic from natural search results and paid ads, including search engine keyword ads, search engine optimization, etc.), social media traffic (traffic from social media platforms, such as social media ads, social media content promotion, etc.), and ad network traffic (traffic from ad networks; recommended ad recipients can place ad codes on ad network platforms to obtain traffic from other websites), etc. The original cost is the pre-set expected cost for the target audience to compete for traffic without using the one-click scaling strategy, before launching the ad campaign. For example, an original cost of 100 means the target audience expects to spend an average of 100 yuan to acquire each click, such as spending 1000 yuan to acquire 10 clicks. The one-click scaling strategy is a scaling product where the target audience allocates a separate scaling budget for their ad. The system will spend this budget within a certain timeframe (usually 6 hours) to acquire more traffic than the original cost could obtain. The scaling budget is the pre-set total amount to be spent on implementing the one-click scaling strategy before launching the ad campaign. For example, a scaling budget of 1000 means that spending 1000 yuan to execute the one-click scaling strategy will acquire more traffic.

[0072] In step 102, in response to receiving the instruction to enable the one-click scaling strategy, the scaling factor corresponding to each piece of recommendation in the recommendation information set is determined based on the recommendation information set, the traffic set, and the scaling budget.

[0073] Here, the scaling factor is used to determine the scaling cost when recommending information competes for traffic. The scaling factor is a specific numerical value, usually greater than 1 (e.g., 1.5). The scaling factor can enhance the competitiveness of recommending information when competing for traffic, helping it acquire the corresponding traffic. When the target audience of the recommended information selects to enable the one-click scaling product on the terminal, the terminal will generate an instruction to enable the one-click scaling strategy and send it to the server. When the server receives the instruction to enable the one-click scaling strategy, it can determine the scaling factor corresponding to each recommending information in the recommended information set based on the recommended information set, the traffic set, and the scaling budget.

[0074] Specifically, see Figure 3B Step 102 can be achieved through steps 1021 to 1023, as explained below:

[0075] In step 1021, the problem to be optimized is constructed based on the set of recommendation information, the set of traffic, and the initial budget.

[0076] Here, the problem to be optimized is defined as maximizing the total incremental conversion amount under the constraint of the initialization budget. The total incremental conversion amount is the total amount of new conversions generated during the recommendation information delivery process through the one-click initialization strategy, that is, the sum of new conversion data such as sales, downloads, and form submissions generated after the recommendation information delivery through the one-click initialization strategy. For example, if the total conversion amount obtained after the recommendation information delivery based on the clicks and orders of the viewers of the recommendation information is 5500 without the one-click initialization strategy enabled, and the total conversion amount obtained after the recommendation information delivery based on the clicks and orders of the viewers of the recommendation information is 10000 without the one-click initialization strategy enabled, then the total incremental conversion amount is 4500. In an exemplary case, the problem to be optimized (i.e., the original problem) can be expressed as formula (1), as follows:

[0077]

[0078] Where Ω represents the set of recommendation information and Ⅱ represents the set of traffic; for any traffic Recommended information random variable x ij Indicates whether recommended information j wins in traffic i, Inc_GMV ij This represents the incremental Gross Merchandise Volume (GMV) of recommended information j under traffic i, where cost is... ij This represents the bidding cost for recommendation information j to compete for traffic i, Budget. j T represents the initial budget for recommending information j. γ This represents the pricing rate (hyperparameter), where the subscript γ can indicate different granularities, such as industry. j is a constraint, representing a flow rate. Under these circumstances, at most one recommendation message will receive this traffic.

[0079] In step 1022, the optimization problem is solved to obtain the optimal initialization cost.

[0080] Here, the optimal initialization cost is the initialization cost that maximizes the total incremental conversion amount. For example, the above formula (1) is transformed into the standard form of linear programming, and the dual problem corresponding to the problem to be optimized is derived. Since the solution to the problem to be optimized and the solution to the dual problem are equivalent when the Karush-Kuhn-Tucker (KKT) conditions are satisfied (the KKT conditions are a set of necessary conditions for judging the optimal solution of a constrained optimization problem, and are an extension of the Lagrange multiplier method), therefore, it is assumed that the solutions to the problem to be optimized and the dual problem are respectively... and Solving equation (1) yields the optimal initial production cost, which can be expressed as equation (2) as follows:

[0081]

[0082] Among them, X ij To maximize the initial cost of incremental conversions. gmv ij This represents the incremental gross merchandise transaction value (gmv) of recommended information j under traffic i. ij It is the estimated value of the incremental gross merchandise transaction volume of recommended information j under traffic i.

[0083] In step 1023, the scaling factor is determined based on the optimal scaling cost and the original cost.

[0084] Here, the ratio between the optimal scaling cost and the original cost is determined as the scaling factor. For example, due to egmv in the above formula (2)... ij Numerically, it satisfies the condition of being equal to the original cost, so the one-click scaling factor can be expressed as:

[0085] In this embodiment, an optimization problem is constructed based on the recommendation information set, traffic set, and initialization budget. The optimal initialization cost is obtained by solving the optimization problem, and then an initialization factor is determined based on the optimal initialization cost and the original cost. In this way, by constructing and solving the optimization problem, the initialization cost that can maximize the incremental conversion amount for the target audience of the recommendation information can be determined. Then, the initialization factor is determined based on the optimal initialization cost and the original cost, improving the efficiency of recommendation message traffic competition and thus improving the delivery efficiency of recommendation messages.

[0086] In step 103, the competition result is obtained by controlling the recommended information set and the competition flow set based on the original cost and the volume factor.

[0087] Here, the cost of recommending information in the recommending information set competing for traffic in the traffic set is determined by the original cost and the starting factor. Then, the traffic is competed for through bidding to obtain the corresponding competition results. During the bidding process, traffic is entered into the bidding queue in order of bid (cost). The recommending information at the top of the bidding queue, i.e., the recommending information with the highest bid, will obtain this traffic. The competition result is used to characterize which recommending information in the traffic set wins each traffic item. For example, the competition result can be represented as: traffic A is obtained by recommending information 1, traffic B is obtained by recommending information 5, and traffic C is obtained by recommending information 4.

[0088] In some embodiments, see Figure 3C Step 103 can be achieved through steps 1031 to 1033, as explained below:

[0089] In step 1031, the first flow and the second flow in the flow set are determined.

[0090] Here, "first-tier traffic" refers to traffic that requires the one-click scaling strategy to be enabled. This means that the recommended information cannot compete for first-tier traffic through the original bid; only by enabling the one-click scaling strategy can it potentially obtain first-tier traffic at a higher scaling cost. "Second-tier traffic" refers to traffic that does not require the one-click scaling strategy to be enabled. This means that the recommended information can compete for second-tier traffic through the original bid and does not need to enable the one-click scaling strategy.

[0091] Specifically, see Figure 3D Step 1031 can be achieved through steps 10311 to 10314, as explained below:

[0092] In step 10311, for each traffic in the traffic set, the bidding queue corresponding to the traffic is determined.

[0093] Here, the bidding queue includes at least one candidate recommendation for competing traffic. The candidate recommendation includes recommendations from the recommendation set, as well as target recommendations not belonging to the recommendation set. The bidding queue refers to the queue of multiple candidate recommendations competing for the same traffic during recommendation message delivery. The candidate recommendations in the bidding queue are sorted by their bids, and the candidate recommendation with the highest bid will receive that traffic.

[0094] In step 10312, the target cost of the competitive traffic for the target recommendation information is determined.

[0095] Here, the target recommendation information refers to candidate recommendation information in the bidding queue that does not belong to the recommendation information set. There can be one or more target recommendation information, and each target recommendation information corresponds to a target cost. The target cost is the bidding cost of the target recommendation information for competing for traffic, which can be determined by the current bid of the target recommendation information competing for traffic in the bidding queue.

[0096] In step 10313, if the original cost is greater than the target cost, the flow rate is determined to be the second flow rate.

[0097] Here, when there is only one target recommendation, there is one target cost. If the original cost is greater than this target cost, the traffic is designated as the second traffic. When there are multiple target recommendations, there are multiple target costs. First, the maximum value of these multiple target costs is determined. If the original cost is greater than this maximum value, then the original cost is greater than the target cost. This means that the bid of the recommendation within the recommendation set is higher than the bid of recommendation not belonging to the recommendation set. Therefore, the recommendation within the recommendation set can obtain this traffic through its original bid. For example, if the original cost is 150, and the target costs include 100, 120, and 70, with the maximum target cost being 120, and the original cost being greater than 120, then the original cost is greater than the target cost, and this traffic is designated as the second traffic.

[0098] In step 10314, if the original cost is less than or equal to the target cost, the flow rate is determined as the first flow rate.

[0099] Similar to step 10313, when there is only one target recommendation, there is one target cost. If the original cost is less than or equal to this target cost, the traffic is designated as the first traffic. When there are multiple target recommendations, there are multiple target costs. First, the maximum value of these multiple target costs is determined. If the original cost is less than or equal to this maximum value, it means the original cost is less than or equal to the target cost. The bid of the recommendation in the recommendation set is not higher than the bid of recommendation not belonging to the recommendation set, so the recommendation in the recommendation set cannot obtain this traffic through its original bid. Therefore, it is necessary to increase the bid of the recommendation to compete for this traffic through a growth factor to help the recommendation obtain traffic. For example, if the original cost is 100, and the target costs include 90, 120, and 130, the largest target cost is 130. The original cost is less than 130, so the original cost is less than the target cost, and therefore this traffic is designated as the first traffic.

[0100] In this embodiment, a bidding queue is determined for each traffic item in the traffic set. The target cost for competing traffic for target recommended information that does not belong to the recommended information set within the bidding queue is determined. If the original cost is greater than the target cost, the traffic is designated as the second traffic item; if the original cost is less than or equal to the target cost, the traffic is designated as the first traffic item. By differentiating the traffic, it is possible to identify which traffic requires a one-click scaling strategy and which does not. This avoids applying the scaling budget to traffic that can be obtained through the original bid, thereby saving the scaling budget and improving the efficiency and profitability of information recommendation.

[0101] In step 1032, for each piece of recommendation information in the recommendation information set, the product of the scaling factor and the original cost is determined as the scaling cost.

[0102] Here, each recommendation has a corresponding growth factor. By multiplying the growth factor of each recommendation by its original cost, we can obtain the growth cost of each recommendation. For example, if the growth factor of recommendation A is 1.2, the growth factor of recommendation B is 1.5, and the original cost is 100, then the growth cost of recommendation A is 120, and the growth cost of recommendation B is 150.

[0103] In step 1033, each recommendation is controlled to compete for the first traffic based on the initial cost and compete for the second traffic based on the original cost, thus obtaining the competition result.

[0104] Here, since the first traffic refers to traffic requiring a one-click scaling strategy, and the second traffic refers to traffic not requiring such a strategy, recommended information will compete for the first traffic based on scaling cost and for the second traffic based on its original cost. Specifically, when competing for the first traffic, recommended information will enter the bidding queue with its scaling cost as the bid, competing with other candidate recommended information in the bidding queue for the first traffic. Similarly, when competing for the second traffic, recommended information will enter the bidding queue with its original cost as the bid, competing with other candidate recommended information in the bidding queue for the second traffic. The competition result is used to characterize which recommended information wins each traffic item in the traffic set. For example, traffic A is won by recommended information 1, and traffic B is won by recommended information 7.

[0105] In this embodiment, a first traffic and a second traffic are determined in the traffic set. For each recommendation in the recommendation information set, the product of the scaling factor and the original cost is determined as the scaling cost. Each recommendation competes for the first traffic based on the scaling cost and for the second traffic based on the original cost, thus obtaining the competition result. In this way, the second traffic that can be obtained through the original cost is not competed for using the scaling cost, while the first traffic that cannot be obtained through the original cost is competed for using the scaling cost. This allows for better utilization of the scaling budget and improves the success rate of information recommendation traffic competition.

[0106] In step 104, based on the competition results, the recommended information set is distributed to the traffic set to obtain the maximum incremental conversion amount.

[0107] Here, the incremental conversion total refers to the total amount of new conversions generated during the recommendation information delivery process through the one-click scaling strategy. This is the sum of new conversion data such as sales revenue, downloads, and form submissions generated after the recommendation information delivery is implemented through the one-click scaling strategy. Maximizing the incremental conversion total ensures that the target audience of the recommendation information receives better returns after the information set is distributed to the traffic set. Based on the competition results, it can be determined which recommendation information won each traffic, and then the recommendation information is delivered to the corresponding traffic. For example, if the competition results are: traffic A is won by recommendation information 1, and traffic B is won by recommendation information 7, then recommendation information 1 will be delivered to traffic A, and recommendation information 7 will be delivered to traffic B.

[0108] In this embodiment, a set of recommended information, a traffic set for delivering the recommended information set, the original cost of delivering the recommended information set to the traffic set, and the initial budget for bidding on the traffic set using a one-click scaling strategy are obtained. Then, upon receiving an instruction to enable the one-click scaling strategy, a scaling factor is determined for each recommended information in the recommended information set based on the recommended information set, the traffic set, and the scaling budget. When bidding on the traffic set for the recommended information set, the cost of bidding for traffic in the traffic set for each recommended information in the recommended information set is controlled based on the original cost and the scaling factor. When the original cost of a recommended information is not competitive enough to acquire traffic through bidding at the original cost, the cost during bidding is increased to the initial cost through the scaling factor, thereby enhancing the competitiveness of the recommended information and helping it acquire traffic. Simultaneously, controlling the scaling factor avoids excessively high initial cost increases that would waste the scaling budget, thus achieving the goal of acquiring more traffic at the lowest possible initial cost within the scaling budget range. This maximizes the total incremental conversion after the recommended information set is delivered to the traffic set, thereby improving the efficiency and profitability of information recommendation.

[0109] In some embodiments, see Figure 3EAfter step 104, the initial budget and original budget can be updated through steps 201 to 204, as explained below:

[0110] In step 201, the original budget for delivering the set of recommended information to the set of traffic is obtained.

[0111] Here, the original budget is the total amount of money that is pre-set for the target audience of the recommended information before it is delivered, which is used for the recommendation information delivery strategy other than one-click scaling. For example, if the original budget is 2000, it means that the total amount of money used for the recommendation information delivery strategy other than one-click scaling is 2000 yuan. You can spend 2000 yuan to deliver the recommended information set to the traffic set without enabling the one-click scaling strategy.

[0112] In step 202, based on the competition results, the incremental flow and existing flow in the flow set are determined.

[0113] Here, incremental traffic refers to the first traffic obtained through competition among recommendation information sets, and stock traffic refers to the second traffic obtained through competition among recommendation information sets. First, based on the competition results, the first and second traffic obtained through competition among recommendation information sets are determined. Then, the first traffic obtained through competition among recommendation information sets is determined as incremental traffic, and the second traffic obtained through competition among recommendation information sets is determined as stock traffic. For example, if the competition results are: traffic A (first traffic) is obtained by recommendation information 1, traffic B (second traffic) is obtained by recommendation information 5, and traffic C (first traffic) is obtained by recommendation information 4; then, the first traffic obtained through competition among recommendation information sets consists of traffic A and traffic C, and the second traffic obtained through competition among recommendation information sets consists of traffic B; therefore, incremental traffic includes traffic A and traffic C, and stock traffic includes traffic B.

[0114] In step 203, for incremental traffic, the scaling budget is updated based on the scaling cost corresponding to at least one first recommendation information.

[0115] Here, the first recommendation information is the recommendation information obtained through competition for incremental traffic. For example, if the competition results are: traffic A (first traffic) is obtained by recommendation information 1, traffic B (second traffic) is obtained by recommendation information 5, and traffic C (first traffic) is obtained by recommendation information 4, then the incremental traffic includes traffic A and traffic C; therefore, the first recommendation information is recommendation information 1 and recommendation information 4. Then, based on the scaling cost and scaling budget corresponding to at least one first recommendation information, the remaining scaling budget is determined and updated to the remaining scaling budget. Specifically, the scaling costs corresponding to at least one first recommendation information can be added together to obtain a sum of multiple scaling costs. Then, the difference between the sum of multiple scaling costs and the scaling budget is determined as the remaining scaling budget, and the scaling budget is updated to the remaining scaling budget. For example, if the scaling costs corresponding to at least one first recommendation information include 130, 150, and 120, and the scaling budget is 500, then the remaining scaling budget is 100, and the scaling budget will be updated to 100. Alternatively, the scaling cost corresponding to at least one first recommendation can be deducted from the scaling budget sequentially to obtain the remaining scaling budget, which is then updated to the remaining scaling budget. For example, if the scaling cost corresponding to at least one first recommendation is 50, 100, or 150, and the scaling budget is 300, then after the first deduction, the remaining scaling budget is 250; after the second deduction, it is 150; and after the third deduction, it is 0, at which point the scaling budget will be updated to 0. Typically, after enabling the one-click scaling strategy, the scaling budget will be fully spent to maximize the total incremental conversion, meaning the remaining scaling budget will be 0.

[0116] In step 204, for the existing traffic, the original budget is updated based on the original cost of at least one second recommendation.

[0117] Here, the second recommendation information refers to the recommendation information obtained from the competition for existing traffic. For example, if the competition results are: traffic A (second traffic) is obtained by recommendation information 1, traffic B (second traffic) is obtained by recommendation information 5, and traffic C (first traffic) is obtained by recommendation information 4, then the existing traffic includes traffic A and traffic B; therefore, the second recommendation information is recommendation information 1 and recommendation information 5. Then, based on the original cost and original budget corresponding to at least one second recommendation information, the remaining original budget is determined and updated to the remaining original budget. Specifically, the original costs corresponding to at least one second recommendation information can be added together to obtain a sum of multiple original costs. The difference between the sum of multiple original costs and the original budget is then determined as the remaining original budget, and the original budget is updated to the remaining original budget. For example, if the original cost is 100, there are 4 second recommendations, and the original budget is 600, then the sum of multiple original costs is 400, and the remaining original budget is 200, which will be updated to 200. Alternatively, the original cost corresponding to at least one second recommendation information can be deducted from the original budget sequentially to obtain the remaining original budget, and the original budget is updated to the remaining original budget. For example, if the original cost is 200, there are 3 second recommendations, and the original budget is 800, then after the first deduction, the remaining original budget is 600, after the second deduction, the remaining original budget is 400, and after the third deduction, the remaining original budget is 200. The original budget will then be updated to 200.

[0118] In this embodiment, the original budget for delivering the recommended information set to the traffic set is obtained. Based on the competition results, the incremental and existing traffic in the traffic set are determined. For incremental traffic, the initial budget is updated based on the initial cost corresponding to at least one first recommended information. For existing traffic, the original budget is updated based on the original cost of at least one second recommended information. This allows traffic obtained through the one-click initialization strategy to be charged according to the initial cost from the initial budget, while traffic obtained through the original cost competition but not through the one-click initialization strategy is charged according to the original cost from the original budget. This saves initial budget, allowing the one-click initialization strategy to play a greater role. The budget is then allocated to other traffic that is difficult to compete for, thereby achieving the goal of obtaining more traffic within the initial budget at the lowest possible initial cost, improving the efficiency and profitability of information recommendation.

[0119] The following will describe an exemplary application of the embodiments of this application in a real-world application scenario, taking an advertising placement scenario as an example.

[0120] In advertising scenarios, recommended information targeting targets have the following needs: 1. Quickly verify the quality of the creative materials themselves, break through bids in a short period to increase traffic, and quickly overcome the cold start period. 2. Expand traffic during major promotional events and holidays: break through bids in a short period to quickly acquire traffic for the ads. The one-click scaling strategy is a scaling product designed to meet the above scenarios. Recommended information targeting targets allocate a separate scaling budget to the ads, and the system aims to spend the scaling budget within a certain period of time, thereby helping recommended information targeting targets quickly acquire traffic, accumulate data for the model, and quickly overcome the cold start period.

[0121] The goal of the one-click scaling strategy is to spend the scaling budget set for the recommended information delivery target within a limited time with the lowest possible scaling cost. To define the business problem of the one-click scaling strategy, and to set the ad set A for which the one-click scaling strategy is enabled, we can obtain formula (3), as follows:

[0122]

[0123] Formula (3) is defined as the problem of maximizing incremental GMV under the constraint of initial budget. The above problem is broken down into a modeling method for traffic. The modeling method is based on two basic assumptions: 1. The traffic is stable and unbiased; 2. The price model (P model) predicts the value without bias. The P model is a model used to predict price changes.

[0124] Under the above assumptions, let the set of recommended information be Ω, and the set of traffic that can participate in bidding each day be II (multiple recommended information may exist under one traffic); for any traffic Recommended information random variable x ij Indicates whether recommended information j wins in traffic i, Inc_GMV ij This represents the incremental GMV of recommended information j under traffic i, and the cost. ij This represents the cost of bidding for ad j at traffic i, Budget j τ represents the one-click initialization budget for recommended information j. γ Let γ represent the pricing rate (hyperparameter), where the subscript γ can represent different granularities, such as industry. The above constrained optimization problem can be written as formula (1):

[0125]

[0126] in, j is a constraint, representing a flow rate. At most, only one recommended message can receive this traffic.

[0127] Rewrite the above formula (1) in standard linear programming form, and derive the dual problem from the primal problem to obtain the dual problem. Since the solutions to the primal and dual problems are equivalent when the KKT conditions are met, assume the solutions to the primal and dual problems are respectively... and Among them, in cost ij Under the dual-price deduction scenario, it equals the effective cost per mille (ECPM) of the second-ranked product. The optimal solution to the optimization problem, i.e., the optimal initial cost, can be expressed as formula (2), as follows:

[0128]

[0129] Therefore, the scaling factor can be determined as follows:

[0130] Then, the scaling factor is applied to the one-click scaling strategy. For ease of understanding, the application of the one-click scaling strategy to ecpm is broken down as follows:

[0131] b*basic_ecpm=(1+b-1)*basic_ecpm

[0132] =basic_ecpm+(b-1)*basic_ecpm

[0133] =original_basic_ecpm+auto_acquisition_basic_ecpm

[0134] Where b represents the accumulator factor, basic_ecpm represents the overall ecpm, original_basic_ecpm represents the ecpm affected by the accumulator factor, and auto_acquisition_basic_ecpm represents the ecpm unaffected by the accumulator factor.

[0135] For each traffic flow, if there are at least one candidate recommendation in the bidding queue competing for that traffic flow, the candidate recommendation in the bidding queue that does not belong to the recommendation set is identified as the target recommendation, and the target cost for the target recommendation to compete for that traffic flow is determined. If the original cost of the recommendation in the recommendation set is greater than the target cost, this traffic flow is identified as the second traffic flow; if the original cost is less than or equal to the target cost, this traffic flow is identified as the first traffic flow.

[0136] When competing for traffic sets using recommended information, the initial cost is determined by the initial cost factor and the original cost. The initial cost is used to compete for the first source of traffic, and this first source is designated as incremental traffic. The original cost is used to compete for the second source of traffic, and this second source is designated as existing traffic. Then, based on the competition results, the recommended information sets are distributed to the traffic sets to maximize the total incremental conversion amount.

[0137] Once the recommended information set is delivered to the traffic set, charges will be deducted. For each piece of traffic obtained through competition for the recommended information set, if the traffic is incremental traffic, the charge will be deducted from the initial cost budget; if the traffic is existing traffic, the charge will be deducted from the original cost budget.

[0138] This application provides two one-click scaling strategies. In one-click scaling strategy (I), traffic is always competed for through the original cost during non-one-click scaling periods, and traffic is always competed for through scaling cost during one-click scaling periods. In one-click scaling strategy (II), traffic is always competed for through the original cost during non-one-click scaling periods. During one-click scaling periods, incremental traffic is obtained by competing for traffic through scaling cost for the first traffic, and existing traffic is obtained by competing for traffic through the original cost for the second traffic.

[0139] See Figure 4 This is a schematic diagram illustrating the effect of the one-click scaling strategy provided in the embodiments of this application. The details are as follows:

[0140] In the one-click scaling strategy (I), traffic is competed for through the original cost during the non-one-click scaling period. Cost A represents all the original costs during the non-one-click scaling period, which are deducted from the original budget. During the one-click scaling strategy, traffic is competed for through scaling costs. Cost B represents all scaling costs during the one-click scaling strategy, which are deducted from the scaling budget.

[0141] In the one-click scaling strategy (II), during the non-one-click scaling period, traffic is competed for through the original cost. Cost A represents all the original costs during the non-one-click scaling period, which are deducted from the original budget. During the one-click scaling period, incremental traffic is obtained by competing for traffic through scaling costs for the first traffic. Cost B1 represents all scaling costs during the one-click scaling period, which are deducted from the scaling budget. During the one-click scaling period, existing traffic is obtained by competing for traffic through the original cost during the one-click scaling period. Cost B2 represents all the original costs during the one-click scaling period, which are deducted from the original budget.

[0142] See Figure 5AThis is a schematic diagram of the non-one-click traffic generation strategy provided in the embodiments of this application. The horizontal axis represents time (H), the vertical axis represents cost, and curve 1 represents the total cost of competing for traffic for recommendation information when the one-click traffic generation strategy is not used.

[0143] See Figure 5B Figure 1 is a schematic diagram of the one-click scaling strategy (I) provided in this application embodiment. The horizontal axis represents time (H), and the vertical axis represents cost. Time T represents the activation time of the one-click scaling strategy (I), and T+n represents the end time of the one-click scaling strategy (I), where n is less than or equal to 6. Curves 1 and 2 represent the total cost of competing for traffic with recommendation information when using the one-click scaling strategy (I). During the one-click scaling strategy (I), the cost is calculated according to curve 2, and the cost corresponding to the dashed part of curve 2 is the scaling cost, which is deducted from the scaling budget. During periods outside the one-click scaling strategy (I), the cost is calculated according to curve 1, and the cost corresponding to the solid part of curve 1 is the original cost, which is deducted from the original budget.

[0144] See Figure 5C Figure 3 is a schematic diagram of the one-click traffic scaling strategy (II) provided in this application embodiment. The horizontal axis represents time (H), and the vertical axis represents cost. Time T represents the activation time of the one-click traffic scaling strategy (II), and T+n represents the end time of the one-click traffic scaling strategy (II), where n is less than or equal to 6. Curves 1 and 3 represent the total cost of competing for recommendation information traffic when using the one-click traffic scaling strategy (II). During periods outside of the one-click traffic scaling strategy (II), costs are calculated according to curve 1, with the cost corresponding to the solid line portion of curve 1 being the original cost, which will be deducted from the original budget. During the one-click traffic scaling strategy (II), costs are calculated according to curve 3, with the incremental cost of the incremental traffic corresponding to the dashed line portion of curve 3 being deducted from the traffic scaling budget, and the original cost of the existing traffic corresponding to the dashed line portion of curve 3 being deducted from the original budget.

[0145] Through the embodiments of this application, when the original cost bidding competitiveness of recommended information is weak and it is impossible to obtain traffic through bidding at the original cost, a scaling factor is used to increase the bidding cost to the scaling cost, thereby enhancing the competitiveness of the recommended information and helping it obtain traffic. Simultaneously, controlling the scaling factor avoids excessively increasing the scaling cost, which would waste the scaling budget. This allows for obtaining more traffic at the lowest possible scaling cost within the scaling budget, maximizing the incremental conversion amount after the recommended information set is deployed to the traffic set, thus improving the efficiency and profitability of information recommendation.

[0146] The following description continues to illustrate the exemplary structure of the information recommendation device 233 provided in the embodiments of this application as a software module. In some embodiments, such as Figure 2 As shown, the information recommendation device 233 stored in the memory 230 may include:

[0147] The acquisition module 2331 is used to acquire a set of recommendation information, a set of traffic for delivering the set of recommendation information, and to acquire the original cost of delivering the set of recommendation information to the set of traffic, as well as the initial budget for the set of recommendation information to bid for the set of traffic through a one-click scaling strategy.

[0148] The determination module 2332 is used to respond to receiving an instruction to enable the one-click scaling strategy, and to determine the scaling factor corresponding to each piece of recommendation in the recommendation information set based on the recommendation information set, the traffic set, and the scaling budget. The scaling factor is used to determine the scaling cost when the recommendation information bids for the traffic set.

[0149] The control module 2333 is used to control the bidding of the recommended information set for the traffic set based on the original cost and the volume factor, so as to obtain the competition result.

[0150] The delivery module 2334 is used to deliver the set of recommended information to the set of traffic based on the competition results, so as to obtain the maximum incremental conversion amount.

[0151] In some embodiments, the determining module 2332 is further configured to construct an optimization problem based on the recommendation information set, the traffic set, and the initialization budget, wherein the optimization problem is defined as maximizing the total incremental conversion amount under the constraint of the initialization budget; solve the optimization problem to obtain the optimal initialization cost, wherein the optimal initialization cost is the initialization cost that can maximize the total incremental conversion amount; and determine the initialization factor based on the optimal initialization cost and the original cost.

[0152] In some embodiments, the control module 2333 is further configured to determine a first flow and a second flow in the flow set, wherein the first flow is the flow that needs to enable the one-click scaling strategy, and the second flow is the flow that does not need to enable the one-click scaling strategy; for each piece of recommendation information in the recommendation information set, the product of the scaling factor and the original cost is determined as the scaling cost; and each piece of recommendation information is controlled to compete for the first flow based on the scaling cost and compete for the second flow based on the original cost to obtain the competition result.

[0153] In some embodiments, the determining module 2332 is further configured to, for each traffic in the traffic set, determine a bidding queue corresponding to the traffic, the bidding queue including at least one candidate recommendation information competing for the traffic; determine the target cost of the target recommendation information competing for the traffic, the target recommendation information being candidate recommendation information in the bidding queue that does not belong to the recommendation information set; if the original cost is greater than the target cost, determine the traffic as the second traffic; if the original cost is less than or equal to the target cost, determine the traffic as the first traffic.

[0154] In some embodiments, the acquisition module 2331 is further configured to acquire the original budget for delivering the set of recommendation information to the set of traffic; based on the competition result, determine the incremental traffic and the existing traffic in the set of traffic, wherein the incremental traffic is the first traffic obtained by the competition for the set of recommendation information, and the existing traffic is the second traffic obtained by the competition for the set of recommendation information; for the incremental traffic, update the initialization budget based on the initialization cost corresponding to at least one first recommendation information, wherein the first recommendation information is the recommendation information obtained by the competition for the incremental traffic; for the existing traffic, update the original budget based on the original cost of at least one second recommendation information, wherein the second recommendation information is the recommendation information obtained by the competition for the existing traffic.

[0155] In some embodiments, the determining module 2332 is further configured to determine, based on the competition result, a first flow and a second flow obtained by the competition of the recommendation information set; determine the first flow obtained by the competition of the recommendation information set as the incremental flow; and determine the second flow obtained by the competition of the recommendation information set as the existing flow.

[0156] In some embodiments, the determining module 2332 is further configured to determine the remaining scaling budget based on the scaling cost corresponding to the at least one first recommendation information and the scaling budget, and update the scaling budget to the remaining scaling budget; and to determine the remaining original budget based on the original cost of the at least one second recommendation information and the original budget, and update the original budget to the remaining original budget.

[0157] This application provides a computer program product comprising a computer program or computer-executable instructions stored in a computer-readable storage medium. The processor of an electronic device reads the computer-executable instructions from the computer-readable storage medium and executes the computer-executable instructions, causing the electronic device to perform the information recommendation method described in this application.

[0158] This application provides a computer-readable storage medium storing computer-executable instructions or a computer program. When the computer-executable instructions or the computer program are executed by a processor, the processor will execute the information recommendation method provided in this application, for example, such as... Figure 3A The information recommendation method is shown.

[0159] In some embodiments, the computer-readable storage medium may be a memory such as RAM, ROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or it may be a variety of devices including one or any combination of the above-mentioned memories.

[0160] In some embodiments, computer-executable instructions may take the form of programs, software, software modules, scripts, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as stand-alone programs or as modules, components, subroutines, or other units suitable for use in a computing environment.

[0161] As an example, computer-executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple co-located files (e.g., files that store one or more modules, subroutines, or code sections).

[0162] As an example, computer-executable instructions can be deployed to execute on a single electronic device, or on multiple electronic devices located at one location, or on multiple electronic devices distributed across multiple locations and interconnected via a communication network.

[0163] In summary, this application's embodiments obtain a set of recommended information, a traffic set for delivering the recommended information set, the original cost of delivering the recommended information set to the traffic set, and the initial budget for bidding on the traffic set using a one-click scaling strategy. Then, upon receiving an instruction to enable the one-click scaling strategy, the scaling factor corresponding to each recommended information in the recommended information set is determined based on the recommended information set, the traffic set, and the scaling budget. When bidding on the traffic set, the cost of each recommended information in the recommended information set is controlled based on the original cost and the scaling factor. When the original cost of a recommended information is weak and cannot acquire traffic through bidding at the original cost, the cost during bidding is increased to the scaling cost using the scaling factor, thereby enhancing the competitiveness of the recommended information and helping it acquire traffic. Simultaneously, it avoids excessively increasing the scaling cost due to the scaling factor, thus avoiding wasting the scaling budget. This achieves the goal of acquiring more traffic at the lowest possible scaling cost within the scaling budget range, maximizing the incremental conversion amount after the recommended information set is delivered to the traffic set, thereby improving the efficiency and profitability of information recommendation.

[0164] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. An information recommendation method, characterized in that, The method includes: Obtain a set of recommended information, a set of traffic for delivering the set of recommended information, and obtain the original cost of delivering the set of recommended information to the set of traffic, as well as the budget for the set of recommended information to compete for the set of traffic through a one-click scaling strategy. In response to receiving an instruction to enable a one-click scaling strategy, based on the recommended information set, the traffic set, and the scaling budget, a scaling factor is determined for each recommended information in the recommended information set. The scaling factor is used to determine the scaling cost when the recommended information competes for the traffic set. Based on the original cost and the scaling factor, the recommended information set competes with the traffic set to obtain the competition result; Based on the competition results, the set of recommended information is distributed to the set of traffic to obtain the maximum total incremental conversion.

2. The method according to claim 1, characterized in that, The step of determining the scaling factor corresponding to each piece of recommendation information in the recommendation information set based on the recommendation information set, the traffic set, and the scaling budget includes: Based on the recommended information set, the traffic set, and the initial volume budget, an optimization problem is constructed. The optimization problem is defined as maximizing the total incremental conversion amount under the constraint of the initial volume budget. Solve the problem to be optimized to obtain the optimal scaling cost, which is the scaling cost that maximizes the total incremental conversion amount; The scaling factor is determined based on the optimal scaling cost and the original cost.

3. The method according to claim 1, characterized in that, The process of controlling the recommendation information set to compete for the traffic set based on the original cost and the scaling factor, and obtaining the competition result, includes: Determine the first traffic and the second traffic in the traffic set, wherein the first traffic is the traffic that needs to enable the one-click traffic boosting strategy, and the second traffic is the traffic that does not need to enable the one-click traffic boosting strategy; For each piece of recommendation information in the set of recommendation information, the product of the scaling factor and the original cost is determined as the scaling cost; Each of the recommended information competes for the first traffic based on the initial cost and competes for the second traffic based on the original cost, thus obtaining the competition result.

4. The method according to claim 3, characterized in that, The determination of the first and second flows in the flow set includes: For each traffic in the traffic set, a bidding queue corresponding to the traffic is determined, and the bidding queue includes at least one candidate recommendation information competing for the traffic; Determine the target cost of competing for the traffic with target recommendation information, wherein the target recommendation information is candidate recommendation information in the bidding queue that does not belong to the recommendation information set; If the original cost is greater than the target cost, the flow rate is determined to be the second flow rate; If the original cost is less than or equal to the target cost, the flow rate is determined to be the first flow rate.

5. The method according to any one of claims 1 to 4, characterized in that, After delivering the recommendation information set to the traffic set based on the competition results, the method further includes: Obtain the original budget for delivering the set of recommended information to the set of traffic; Based on the competition results, incremental traffic and existing traffic in the traffic set are determined. The incremental traffic is the first traffic obtained from the competition of the recommendation information set, and the existing traffic is the second traffic obtained from the competition of the recommendation information set. For the incremental traffic, the scaling budget is updated based on the scaling cost corresponding to at least one first recommendation information, where the first recommendation information is the recommendation information for the incremental traffic obtained through competition. For the existing traffic, the original budget is updated based on the original cost of at least one second recommendation information, where the second recommendation information is the recommendation information obtained through competition for the existing traffic.

6. The method according to claim 5, characterized in that, The step of determining the incremental traffic and existing traffic in the traffic set based on the competition results includes: Based on the competition results, determine the first flow and the second flow obtained from the competition for the recommended information set; The first flow obtained from the competition of the recommended information set is determined as the incremental flow; The second flow obtained by competing for the recommended information set is determined as the existing flow.

7. The method according to claim 5, characterized in that, The step of updating the scaling budget based on the scaling cost corresponding to at least one first recommendation information includes: Based on the initial cost corresponding to the at least one first recommendation information and the initial budget, the remaining initial budget is determined, and the initial budget is updated to the remaining initial budget. The process of updating the original budget based on the original cost of at least one second recommendation includes: Based on the original cost of the at least one second recommendation information and the original budget, the remaining original budget is determined, and the original budget is updated to the remaining original budget.

8. An information recommendation device, characterized in that, The device includes: The acquisition module is used to acquire a set of recommendation information, a set of traffic for delivering the set of recommendation information, the original cost of delivering the set of recommendation information to the set of traffic, and the initial budget for the set of recommendation information to bid for the set of traffic through a one-click scaling strategy. The determination module is used to respond to receiving an instruction to enable the one-click scaling strategy, and to determine the scaling factor corresponding to each piece of recommendation in the recommendation information set based on the recommendation information set, the traffic set, and the scaling budget. The scaling factor is used to determine the scaling cost when the recommendation information bids for the traffic set. The control module is used to control the bidding of the recommended information set for the traffic set based on the original cost and the volume factor, so as to obtain the competition result; The delivery module is used to deliver the set of recommended information to the set of traffic based on the competition results, so as to obtain the maximum incremental conversion amount.

9. An electronic device, characterized in that, The electronic device includes: Memory is used to store executable instructions for a computer; A processor, when executing computer-executable instructions stored in the memory, implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing computer-executable instructions or a computer program, characterized in that, When the computer-executable instructions or computer program are executed by a processor, they implement the method described in any one of claims 1 to 7.

11. A computer program product comprising computer-executable instructions or a computer program, characterized in that, When the computer-executable instructions or computer program are executed by a processor, they implement the method according to any one of claims 1 to 7.

Images