Advertisement putting strategy optimization method and device and storage medium

Abstract

The invention provides an advertisement putting strategy optimization method and device and a storage medium, and relates to the field of advertisement flow putting, and the method comprises the following steps: constructing a virtual advertisement network: generating a virtual advertisement which is consistent with a main advertisement material in vision, but is lower than the main advertisement in commercial value, and the virtual advertisement is directionally put into the keyword bidding pool which is the same as the main advertisement. The virtual advertisement is set to attract the robot to click, and the probability that the robot clicks the main advertisement is reduced, so that the advertisement fee budget consumption of the main advertisement can be reduced, the influence of unreal data such as robot clicking on the main advertisement flow data can be reduced, and people can conveniently select the main advertisement according to the flow data of the main advertisement. The advertisement content is adjusted, the advertisement quality and conversion rate are improved, the robot clicks the virtual advertisement, information of more robots can also be collected, subsequent blacklist pulling is facilitated, the influence of the subsequent robot on the main advertisement is reduced, the budget loss is reduced, and the advertisement conversion rate is improved.

Description

Technical Field

[0001] The present invention relates to the field of advertising traffic delivery, and in particular to an optimization method, device and storage medium for an advertising delivery strategy. Background Art

[0002] In the existing e-commerce advertising bidding, competitors may use robots to click on their opponents' ads, quickly consume their budgets, and cause their ads to be offline early, thereby allowing their own ads to gain more exposure opportunities. Advertising systems usually have anti-fraud mechanisms, such as detecting abnormal click frequencies, IP addresses, device IDs, etc., but advanced robots may use distributed IPs and simulate real user behaviors to bypass detection, resulting in the consumption of budgets. When the advertiser's budget reaches the daily threshold, the ad will stop displaying. When this happens, it will affect the conversion rate of its own ads.

[0003] Therefore, we have made improvements to this and proposed an optimization method, device and storage medium for advertising delivery strategy. Summary of the invention

[0004] The purpose of the present invention is to address the problem that currently existing advanced robots may use distributed IPs and simulate real user behaviors to bypass detection, resulting in the consumption of advertising budgets. When the advertising budget of an advertiser reaches the daily threshold, the advertisements will stop being displayed. When this happens, the conversion rate of the advertisements themselves will be affected.

[0005] In order to achieve the above-mentioned purpose of the invention, the present invention provides an optimization method, device and storage medium for advertising delivery strategy to improve the above-mentioned problem.

[0006] The specific application is as follows: The following steps are involved: S100. Build a virtual advertising network: Generate virtual ads that are visually consistent with the main ad material but have lower commercial value than the main ad, and direct the virtual ads to the same keyword bidding pool as the main ad (placing virtual ads that are visually consistent with the main ad material in the same keyword bidding pool as the main ad can not only be used to attract robot attacks, but also attract real consumers to click on the main ad, thereby improving publicity effects and conversion rates, as the ad content of the virtual ad is the same as the main ad but the product price is different, which has a price advantage), and the bid of the main ad in the ad space bidding is 5% higher than the bid of the virtual ad, the advertising budget of the main ad is 90% of the total advertising budget, and the advertising budget of the virtual ad is 10% of the total advertising budget; S200, collect behavior data: collect user click behavior data on virtual advertisements and main advertisements in real time, including: user IP address, click timestamp, click frequency, page dwell time and device fingerprint (device fingerprint refers to a device feature or unique device identifier that can be used to uniquely identify the device. The collection of device fingerprints is an existing well-known technology, so it is not repeated here. By collecting device fingerprints, users who click on virtual advertisements can be blacklisted by the method in S600 during robot high-frequency periods, so that they cannot click on main advertisements in the future). (Since the e-commerce platform has a robot detection mechanism to prevent malicious competition, traditional robots stay on the advertisement page for a fixed time after clicking on the advertisement, but advanced robots are more random, so the page dwell time is not fixed. Therefore, it is necessary to collect the page dwell time. Combined with S300, when the click frequency of the same IP address exceeds the threshold N times / minute and the page dwell time is less than T seconds, it is determined that the current period is a robot click high-frequency period, and the error is relatively small); S300, detecting robot high-frequency time period: when the virtual advertisement and the main advertisement detect that the click frequency of the same IP address exceeds the frequency threshold N times / minute within the set time window, and the page stay time is less than the time length T seconds, it is determined that the current is a robot high-frequency click period; S400, Dynamic Budget Allocation and Priority Control: S401. During robot-frequent periods, the proportion of virtual advertisements' advertising budget is increased to 10%-20% of the total advertising budget (the total advertising budget is the sum of the virtual advertisements' advertising budget and the main advertisements' advertising budget). By increasing the virtual advertisements' advertising budget, they can compete for the 1st to 3rd positions, attract robot clicks, and reduce the probability of robots clicking on the main advertisements. This can not only reduce the consumption of the main advertisements' advertising budget, but also reduce the impact of unreal data such as robot clicks on the main advertisements' traffic data, making it easier for people to adjust advertisement content based on the main advertisements' traffic data, improve advertisement quality and conversion rate, increase the virtual advertisements' advertising budget to attract robot attacks, and collect more robot information for subsequent blacklisting, thereby reducing the impact of subsequent robots on the main advertisements, reducing budget losses, and improving advertisement conversion rates. S402, synchronously reduce the advertising budget of the main advertisement by 10%-20%, so that the main advertisement position is controlled at 4-6th place, so that the robot can give priority to clicking virtual advertisements and other high-ranking advertisements, reducing the loss of the advertising budget of the main advertisement. In addition, the robot can give priority to clicking virtual advertisements and other high-ranking advertisements. First, when the advertising budget of the virtual advertisement is exhausted, the main advertisement can immediately stop bidding and adjust, reducing the loss of the advertising budget of the main advertisement during the robot's click peak. Second, when the advertising budget of the virtual advertisement is exhausted, it will withdraw from the advertising position competition, which can improve the advertising position of the main advertisement. Therefore, there is no need to adjust the bidding of the main advertisement, and the advertising position can be improved. In addition, when the advertising budget of the virtual advertisement is exhausted, the advertising budget of other high-ranking advertisements will be consumed first. When the advertising budget of other high-ranking advertisements is exhausted, the advertising position of the main advertisement can also be improved, and the resources of the attacking party can also be consumed. S403: When the advertising budget of the virtual advertisement is exhausted, the main advertisement immediately stops the bid adjustment and maintains the current ranking; S500, bidding strategy adjustment: S501. During the robot high-frequency period, the bid for the virtual advertisement is increased by 10%-20% to improve the ranking of the virtual advertisement, and the bid for the main advertisement is reduced by 10%-20% to reduce the ranking of the main advertisement; S502: When no high-frequency time period features of a robot are detected for M consecutive minutes, the bid for the main advertisement is increased by 40%-60%.

[0007] As the preferred technical solution of this application, the specific rules for budget allocation in S400 are: A. The total advertising budget is fixed at B, and the initial advertising budget for virtual advertising accounts for 10%; B. When the robot's high-frequency period is triggered, the virtual advertising budget is adjusted to min(0.2B, B-main advertising budget consumed); C. The cost per click of virtual advertising is fixed at C virtual , the maximum number of clicks is K = virtual advertising budget / C virtual , K is rounded down to an integer; D. When the number of clicks on the virtual ad reaches K times, the ad space of the virtual ad is automatically closed.

[0008] As a preferred technical solution of this application, the threshold value in S300 is set as: Frequency threshold N = twice the highest click frequency of normal users in history, and no less than 10 times / minute; Duration threshold T = 20% of the average duration of the main advertisement, and not more than 5 seconds.

[0009] As a preferred technical solution of the present application, the advertisement sequence control logic in S400 includes: Virtual advertising bid meets: Bid virtual ≥1.2 times the bid of the third-ranked ad slot; Main ad bid adjustment: Bid main =Current bid of primary ad×[0.4+0.1×(6-target ranking)], where the target ranking is 4-6.

[0010] As a preferred technical solution of the present application, S500 also includes a budget reallocation mechanism: When the number of clicks reaches K and the virtual ad is closed, 90% of the remaining budget of the virtual ad will be immediately transferred to the main ad's budget. After receiving the transferred advertising budget, the bidding formula of the main ad is adjusted to: Bid main-new =Bid main ×[1+transferred advertising budget / (0.5×primary advertising original budget)]; This bid will last until the advertising budget is exhausted.

[0011] As a preferred technical solution of the present application, the value of M in S500 is: M=max(15 minutes, number of robot sessions detected in the current period × 60 minutes / total number of sessions).

[0012] As the preferred technical solution of this application, it also includes: ‌S600, Abnormal Traffic Marking and Troubleshooting‌: For users who click on virtual ads during robot-frequent periods, their device fingerprints are marked and added to the blacklist; When a device in the blacklist requests an ad again, only a dummy ad is returned and the click is not counted as billable; The validity period of the blacklist is dynamically set according to the risk level, ranging from 1 hour to 7 days; Blacklist validity period formula: ,in Indicates the number of times users click on virtual ads during robot high frequency periods.

[0013] As the preferred technical solution of this application, it also includes: ‌S700, Risk Period Attenuation Model‌: Calculate the robot click ratio R = robot clicks / total clicks every 5 minutes; When R decreases for three consecutive periods and is lower than 10%, it is determined to exit the risk period; The main ad bid is restored in a gradient manner: the current main ad bid is increased by 20% each cycle until it returns to the original bid.

[0014] An optimization device for advertising delivery strategy includes the following modules: Advertisement delivery module: used to deliver virtual advertisements and main advertisements to the bidding pool that matches the keywords of the main advertisements; Data collection module: used to collect users’ click behavior data on virtual ads and main ads in real time; Detection module: used to detect the high-frequency click period of the robot in the virtual advertisement and the main advertisement within the set time window; Dynamic budget allocation module: adjust the advertising budget ratio of virtual ads and main ads during the high-frequency period of robots; Bidding strategy adjustment module: adjust the bids for virtual ads and main ads during high-frequency periods of robots. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of an optimization method for an advertisement delivery strategy provided in this application; Figure 2 A schematic diagram of an optimization device for an advertisement delivery strategy provided in this application. DETAILED DESCRIPTION

[0016] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0017] As described in the background technology, advanced robots may use distributed IPs and simulate real user behaviors to bypass detection, resulting in the consumption of advertising budgets. When the advertiser's advertising budget reaches the daily threshold, the advertisement will stop being displayed. When this happens, the conversion rate of the advertisement itself will be affected.

[0018] In order to solve this technical problem, the present invention provides an optimization method, device and storage medium for an advertisement delivery strategy, which are applied to advertisement traffic delivery.

[0019] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings.

[0020] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features and technical solutions in the embodiments may be combined with each other.

[0021] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0022] Example 1, please refer to Figure 1 , an optimization method for an advertisement delivery strategy, comprising the following steps: S100. Build a virtual advertising network: Generate virtual ads that are visually consistent with the main ad material but have lower commercial value than the main ad (low commercial value is mainly reflected in the fact that the price of the advertised product is 50% or more higher than the price of the main ad product and has no actual transaction value), and direct the virtual ads to the same keyword bidding pool as the main ad (putting virtual ads that are visually consistent with the main ad material into the same keyword bidding pool as the main ad can not only be used to attract robot attacks, but also attract real consumers to click on the main ad, thereby improving publicity effects and conversion rates, since the ad content of the virtual ad is the same as the main ad but the product price is different, there is a price advantage), and the bid of the main ad in the ad space bidding is 5% higher than the bid of the virtual ad, the advertising budget of the main ad is 90% of the total advertising budget, and the advertising budget of the virtual ad is 10% of the total advertising budget (the bid of the main ad in the ad space bidding is 5% higher than the bid of the virtual ad, ensuring that the ad space ranking of the main ad is higher than the virtual ad at this time, and more of the advertising budget is used for the promotion of the main ad); S200, collect behavior data: collect user click behavior data on virtual advertisements and main advertisements in real time, including: user IP address, click timestamp, click frequency, page dwell time and device fingerprint (device fingerprint refers to a device feature or unique device identifier that can be used to uniquely identify the device. The collection of device fingerprints is an existing well-known technology, so it is not repeated here. By collecting device fingerprints, users who click on virtual advertisements can be blacklisted by the method in S600 during robot high-frequency periods, so that they cannot click on main advertisements in the future). (Since the e-commerce platform has a robot detection mechanism to prevent malicious competition, traditional robots stay on the advertisement page for a fixed time after clicking on the advertisement, but advanced robots are more random, so the page dwell time is not fixed. Therefore, it is necessary to collect the page dwell time. Combined with S300, when the click frequency of the same IP address exceeds the threshold N times / minute and the page dwell time is less than T seconds, it is determined that the current period is a robot click high-frequency period, and the error is relatively small); S300, detecting robot high-frequency time period: when the virtual advertisement and the main advertisement detect that the click frequency of the same IP address exceeds the frequency threshold N times / minute within the set time window, and the page stay time is less than the time length T seconds, it is determined that the current is a robot high-frequency click period; S400, Dynamic Budget Allocation and Priority Control: S401. During robot-frequent periods, the proportion of virtual advertisements' advertising budget is increased to 10%-20% of the total advertising budget (the total advertising budget is the sum of the virtual advertisements' advertising budget and the main advertisements' advertising budget). By increasing the virtual advertisements' advertising budget, they can compete for the 1st to 3rd positions, attract robot clicks, and reduce the probability of robots clicking on the main advertisements. This not only reduces the consumption of the main advertisements' advertising budget, but also reduces the impact of unreal data such as robot clicks on the main advertisements' traffic data, making it easier for people to adjust advertisement content based on the main advertisements' traffic data, improve advertisement quality and conversion rate, increase the virtual advertisements' advertising budget to attract robot attacks, and collect more robot information, making it easier to blacklist them later, reducing the impact of subsequent robots on the main advertisements. S402, synchronously reduce the advertising budget of the main advertisement by 10%-20%, so that the main advertisement position is controlled at 4-6th place, so that the robot can give priority to clicking virtual advertisements and other high-ranking advertisements, reducing the loss of the advertising budget of the main advertisement. In addition, the robot can give priority to clicking virtual advertisements and other high-ranking advertisements. First, when the advertising budget of the virtual advertisement is exhausted, the main advertisement can immediately stop bidding and adjust, reducing the loss of the advertising budget of the main advertisement during the robot's click peak. Second, when the advertising budget of the virtual advertisement is exhausted, it will withdraw from the advertising position competition, which can improve the advertising position of the main advertisement. Therefore, there is no need to adjust the bidding of the main advertisement, and the advertising position can be improved. In addition, when the advertising budget of the virtual advertisement is exhausted, the advertising budget of other high-ranking advertisements will be consumed first. When the advertising budget of other high-ranking advertisements is exhausted, the advertising position of the main advertisement can also be improved, and the resources of the attacking party can also be consumed. S403: When the advertising budget of the virtual advertisement is exhausted, the main advertisement immediately stops the bid adjustment and maintains the current ranking; S500, bidding strategy adjustment: S501. During the robot high-frequency period, the bid for the virtual advertisement is increased by 10%-20% to improve the ranking of the virtual advertisement, and the bid for the main advertisement is reduced by 10%-20% to reduce the ranking of the main advertisement; S502. When no robot high-frequency time period features are detected for M consecutive minutes, the bid for the main advertisement is increased by 40%-60% to compete for the 1st to 3rd ranking advertisement positions.

[0023] Example 2 further optimizes the optimization method of the advertising delivery strategy provided in Example 1. Specifically, Figure 1 As shown, the specific rules for budget allocation in S400 are: A. The total advertising budget is fixed at B, and the initial advertising budget for virtual advertising accounts for 10%; B. When the robot's high-frequency period is triggered, the virtual advertising budget is adjusted to min(0.2B, B-main advertising budget consumed); C. The cost per click of virtual advertising is fixed at C virtual , the maximum number of clicks is K = virtual advertising budget / C virtual , K is rounded down to an integer; D. When the number of clicks on the virtual ad reaches K times, the ad space of the virtual ad is automatically closed.

[0024] The thresholds in S300 are set as: Frequency threshold N = twice the highest click frequency of normal users in history, and no less than 10 times / minute; Duration threshold T = 20% of the average duration of the main advertisement, and not more than 5 seconds.

[0025] The advertisement sequence control logic in S400 includes: Virtual advertising bid meets: Bid virtual ≥1.2 times the bid of the third-ranked ad slot; Main ad bid adjustment: Bid main =Current bid of main ad×[0.4+0.1×(6-target ranking)], where the target ranking is 4-6; The advertising exchange platform updates the bid ranking every 5 seconds to ensure that the main advertisement is maintained in the 4-6th position in real time.

[0026] The S500 also includes a budget reallocation mechanism: When the number of clicks reaches K and the virtual ad is closed, 90% of the remaining budget of the virtual ad's advertising fee is immediately transferred to the main ad's advertising fee budget. Considering the delay in deduction by the advertising platform and because K is rounded down to an integer, the actual consumption may be slightly lower than the budget cap, resulting in a surplus. After receiving the transferred advertising budget, the bidding formula of the main ad is adjusted to: Bid main-new =Bid main ×[1+transferred advertising budget / (0.5×primary advertising budget)], which prevents robots from over-consuming the budget and maximizes the efficiency of the remaining budget; This bid will last until the advertising budget is exhausted.

[0027] The value of M in S500 is: M=max(15 minutes, number of robot sessions detected in the current period × 60 minutes / total number of sessions).

[0028] The device fingerprint hash value is generated as follows: Hardware parameters of acquisition equipment: screen resolution, GPU model, battery capacity; Collection environment parameters: operating system version hash, browser plug-in list MD5 value; Generate a string by concatenating the above parameters and use the SHA-256 algorithm to generate a unique device fingerprint; Parameter precedence rules: Hardware parameter weights: screen resolution (30%), GPU model (25%), battery capacity (20%); Software parameter weights: operating system version hash (15%), browser plug-in list MD5 (10%); Collision handling: When hash values ​​conflict, enable alternate parameters (such as device memory size, number of CPU cores) for recalculation; Also includes: ‌S600, Abnormal Traffic Marking and Troubleshooting‌: For users who click on virtual ads during robot-frequent periods, their device fingerprints are marked and added to the blacklist; When a device in the blacklist requests an ad again, only a dummy ad is returned and the click is not counted as billable; The validity period of the blacklist is dynamically set according to the risk level, ranging from 1 hour to 7 days; Blacklist validity period formula: ,in Indicates the number of times users click on virtual ads during robot high frequency periods.

[0029] Also includes: ‌S700, Risk Period Attenuation Model‌: Calculate the robot click ratio R = robot clicks / total clicks every 5 minutes; When R decreases for three consecutive periods and is lower than 10%, it is determined to exit the risk period; The main ad bid is restored in a gradient manner: the current main ad bid is increased by 20% each cycle until it returns to the original bid.

[0030] Example 3, please refer to Figure 2 , an optimization device for advertising delivery strategy, comprising the following modules: Advertisement delivery module: used to deliver virtual advertisements and main advertisements to the bidding pool that matches the keywords of the main advertisements; Data collection module: used to collect users’ click behavior data on virtual advertisements and main advertisements in real time; Detection module: used to detect the high-frequency click period of the robot in the virtual advertisement and the main advertisement within the set time window; Dynamic budget allocation module: adjust the advertising budget ratio of virtual ads and main ads during the high-frequency period of robots; Bidding strategy adjustment module: adjust the bids for virtual ads and main ads during high-frequency periods of robots.

[0031] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral one; it can be a mechanical connection, an electrical connection, or communication with each other; it can be a direct connection, or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0032] Obviously, the embodiments described above are only some embodiments of the present invention, rather than all embodiments. The preferred embodiments of the present invention are given in the accompanying drawings, but they do not limit the patent scope of the present invention. The present invention can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present invention more thorough and comprehensive. Although the present invention has been described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to replace some of the technical features therein with equivalents. Any equivalent structure made using the contents of the specification and drawings of the present invention, directly or indirectly used in other related technical fields, is also within the scope of patent protection of the present invention.

Claims

1. A method for optimizing an advertising delivery strategy, characterized in that: The following steps are involved: S100. Build a virtual advertising network: Generate virtual ads that are visually consistent with the main ads but have lower commercial value than the main ads, and direct the virtual ads to the same keyword bidding pool as the main ads. The bid of the main ads in the ad space bidding is 5% higher than the bid of the virtual ads. The advertising budget of the main ads is 90% of the total advertising budget, and the advertising budget of the virtual ads is 10% of the total advertising budget. S200, collecting behavior data: collecting the user's click behavior data on the virtual advertisement and the main advertisement in real time, including: user IP address, click timestamp, click frequency, page dwell time and device fingerprint; S300, detecting robot high-frequency time period: when the virtual advertisement and the main advertisement detect that the click frequency of the same IP address exceeds the frequency threshold N times / minute within the set time window, and the page stay time is less than the time threshold T seconds, it is determined that the current is a robot high-frequency click period; S400, Dynamic Budget Allocation and Priority Control: S401. During robot-intensive periods, increase the proportion of virtual advertising budget to 10%-20% of the total advertising budget; S402. Simultaneously reduce the advertising budget of the main advertisement by 10%-20%; S403: When the advertising budget of the virtual advertisement is exhausted, the main advertisement immediately stops the bid adjustment and maintains the current ranking; S500, bidding strategy adjustment: S501. During the robot high-frequency period, the bid for the virtual advertisement is increased by 10%-20% to improve the ranking of the virtual advertisement, and the bid for the main advertisement is reduced by 10%-20% to reduce the ranking of the main advertisement; S502: When no high-frequency time period features of a robot are detected for M consecutive minutes, the bid for the main advertisement is increased by 40%-60%.

2. The method for optimizing an advertisement delivery strategy according to claim 1, characterized in that: The specific rules for budget allocation in S400 are: A. The total advertising budget is fixed at B, and the initial advertising budget for virtual advertising accounts for 10%; B. When the robot's high-frequency period is triggered, the virtual advertising budget is adjusted to min(0.2B, B-main advertising budget consumed); C. The cost per click of virtual advertising is fixed at C virtual , the maximum number of clicks is K = virtual advertising budget / C virtual , K is rounded down to an integer; D. When the number of clicks on the virtual ad reaches K times, the ad space of the virtual ad is automatically closed.

3. The method for optimizing an advertisement delivery strategy according to claim 2, characterized in that: The thresholds in S300 are set as: Frequency threshold N = twice the highest click frequency of normal users in history, and no less than 10 times / minute; Duration threshold T = 20% of the average duration of the main advertisement, and not more than 5 seconds.

4. The method for optimizing an advertisement delivery strategy according to claim 3, characterized in that: The advertisement sequence control logic in S400 includes: Virtual advertising bid meets: Bid virtual ≥1.2 times the bid of the third-ranked ad slot; The main ad bid is adjusted to: Bid main =Current bid of primary ad×[0.4+0.1×(6-target ranking)], where the target ranking is 4-6.

5. The method for optimizing an advertisement delivery strategy according to claim 4, characterized in that: The S500 also includes a budget reallocation mechanism: When the number of clicks reaches K and the virtual ad is closed, 90% of the remaining budget of the virtual ad will be immediately transferred to the main ad's budget. After receiving the transferred advertising budget, the bidding formula of the main ad is adjusted to: Bid main-new =Bid main ×[1+transferred advertising budget / (0.5×primary advertising original budget)]; This bid will last until the advertising budget is exhausted.

6. The method for optimizing an advertisement delivery strategy according to claim 5, characterized in that: The value of M in S500 is: M=max(15 minutes, number of robot sessions detected in the current period × 60 minutes / total number of sessions).

7. The method for optimizing an advertisement delivery strategy according to claim 6, characterized in that: Also includes: ‌S600, Abnormal Traffic Marking and Troubleshooting‌: For users who click on virtual ads during robot-frequent periods, their device fingerprints are marked and added to the blacklist; When a device in the blacklist requests an ad again, only a dummy ad is returned and the click is not counted as billable; The validity period of the blacklist is dynamically set according to the risk level, ranging from 1 hour to 7 days; Blacklist validity period formula: ,in Indicates the number of times users click on virtual ads during robot high frequency periods.

8. The method for optimizing an advertisement delivery strategy according to claim 7, characterized in that: Also includes: ‌S700, Risk Period Attenuation Model‌: Calculate the robot click ratio R = robot clicks / total clicks every 5 minutes; When R decreases for three consecutive periods and is lower than 10%, it is determined to exit the risk period; The main ad bid is restored in a gradient manner: the current main ad bid is increased by 20% each cycle until it returns to the original bid.

9. An optimization device for advertising delivery strategy, using the optimization method for advertising delivery strategy as claimed in claim 8, characterized in that: Includes the following modules: Advertisement delivery module: used to deliver virtual advertisements and main advertisements to the bidding pool that matches the keywords of the main advertisements; Data collection module: used to collect users’ click behavior data on virtual advertisements and main advertisements in real time; Detection module: used to detect the high-frequency click period of the robot in the virtual advertisement and the main advertisement within the set time window; Dynamic budget allocation module: adjust the advertising budget ratio of virtual ads and main ads during the high-frequency period of robots; Bidding strategy adjustment module: adjust the bids for virtual ads and main ads during high-frequency periods of robots.

10. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images