Advertisement putting strategy optimization method and system based on AI learning

Abstract

The embodiment of the invention relates to the technical field of artificial intelligence, and discloses an advertisement putting strategy optimization method and system based on AI learning, and the method comprises the steps: firstly, capturing real-time state parameters of an external environment through a multi-source environment sensor, and analyzing and associating hotspot online shopping data in combination with big data; quantifying the matching degree between each candidate advertisement and the environment state through a pre-training model, and generating a first putting strategy; and then, in an advertisement display process, performing secondary optimization on the first putting strategy to obtain a second putting strategy by taking recessive behavior data such as user watching duration and the like as feedback indexes. Therefore, by combining the environment parameters and the online shopping hotspot data and utilizing the model score sorting to optimize the initial strategy, the advertisement putting better fits the environment state and the online shopping consumption trend, the adaptability of the advertisement putting with the scene environment and the online shopping hotspot is greatly improved, the advertisement putting effect and the user attention are improved, and the advertisement putting experience is improved. And intelligent and precise optimization of an advertisement putting strategy is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence, and particularly to an optimization method and system for advertising placement strategies based on AI learning. Background Art

[0002] In the era of digital marketing, the accuracy and efficiency of advertising placement are directly related to the operational benefits of business entities. Traditional shopping mall advertising placement strategies usually arrange content based on fixed time periods, preset rules, or manual experience, and it is difficult to dynamically adapt to environmental changes and real-time fluctuations in user behavior. For example, seasonal changes, weather changes, or emergencies may cause significant changes in consumers' shopping preferences, while static advertising strategies cannot capture such environmental correlations in a timely manner, resulting in waste of resources and low conversion rates. At the same time, the booming development of online consumer behavior has made users' shopping decisions show the characteristics of "integration of online and offline", but offline advertising systems lack the ability to integrate real-time online shopping trends, resulting in insufficient matching between advertising content and users' potential needs. Summary of the Invention

[0003] The main objective of the present invention is to provide an optimization method and system for advertising placement strategies based on AI learning, aiming to solve the technical problem in the prior art that the matching between advertising content and users' potential needs is insufficient due to static advertising placement.

[0004] To achieve the above objective, in a first aspect, an optimization method for advertising placement strategies based on AI learning is provided in an embodiment of the present application, which is applied to an advertising placement strategy optimization system. The advertising placement strategy optimization system includes an environmental sensor and an advertising display screen. The method includes: Obtain environmental status parameters collected in real time by the environmental sensor, where the environmental status parameters include at least one of indoor environmental parameters and outdoor environmental parameters; Obtain hot online shopping data that matches the environmental status parameters according to the environmental status parameters; Input the environmental status parameters and the hot online shopping data into a pre-trained advertising matching model, and output the matching degree scores of each candidate advertisement; Rank the priority of each candidate advertisement according to the matching degree scores, and generate a first advertising placement strategy to preliminarily optimize the initial advertising placement strategy; Control the playing content of the advertising display screen according to the first advertising placement strategy, and collect user gaze duration data in real time during the playing process; Re-optimize the first advertising placement strategy according to the user gaze duration data to obtain a second advertising placement strategy.

[0005] In a possible implementation, the environmental sensor includes an indoor environmental sensor and an outdoor environmental sensor, and obtains environmental state parameters collected in real time by the environmental sensor, including: Obtain the indoor environmental state parameters collected in real time by the indoor environmental sensor, where the indoor environmental state parameters at least include the indoor population density; Obtain the outdoor environmental state parameters collected in real time by the outdoor environmental sensor, where the outdoor environmental state parameters at least include one of the environmental temperature and the environmental rainfall.

[0006] In a possible implementation, obtain hot online shopping data that matches the environmental state parameters according to the environmental state parameters, including: When the indoor population density is greater than or equal to the population density threshold and the environmental temperature is greater than or equal to the temperature threshold, use a reinforcement learning search algorithm to obtain the hot-selling product data of online shopping in hot weather; Or, when the indoor population density is greater than or equal to the population density threshold and the environmental rainfall is greater than or equal to the rainfall threshold, use a reinforcement learning search algorithm to obtain the hot-selling product data of online shopping in rainy weather.

[0007] In a possible implementation, input the environmental state parameters and the hot online shopping data into a pre-trained advertisement matching model, and output the matching degree scores of each candidate advertisement, including: Input the environmental state parameters into the environmental sub-network of the pre-trained advertisement matching model to obtain the environmental attribute scores of each candidate advertisement; Input the hot online shopping data into the hot sub-network of the pre-trained advertisement matching model to obtain the hot attribute scores of each candidate advertisement; Perform weighted summation on the environmental attribute scores and the hot attribute scores to obtain the matching degree scores of each candidate advertisement.

[0008] In a possible implementation, before performing weighted summation on the environmental attribute scores and the hot attribute scores to obtain the matching degree scores of each candidate advertisement, it further includes: Obtain the change rate of the environmental state parameters and the trend acceleration of the hot online shopping data; Input the change rate of the environmental state parameters and the trend acceleration of the hot online shopping data into a dynamic selection network to generate dynamic weight coefficients, where the dynamic selection network satisfies the following expression: ;

[0009] ;

[0010] ;

[0011] In the formula, is the dynamic weight for environmental attribute scoring, is the dynamic weight for hot-spot attribute scoring; is the change rate of environmental state parameters, is the reference change rate of environmental state parameters, is the reference weight for environmental attribute scoring; is the trend acceleration of hot-spot online shopping data, is the reference trend acceleration of hot-spot online shopping data, is the reference weight for hot-spot attribute scoring; is the warning value of the change rate of environmental state parameters, is the warning value of the trend acceleration of hot-spot online shopping data.

[0012] In a possible implementation, the candidate advertisements are prioritized according to the matching degree score, and a first advertisement placement strategy is generated to preliminarily optimize the initial advertisement placement strategy, including: The candidate advertisements are prioritized according to the descending order of the matching degree score to obtain an advertisement placement playlist; The display duration of each candidate advertisement in the advertisement placement playlist is determined according to the matching degree score of each candidate advertisement.

[0013] In a possible implementation, determining the display duration of each candidate advertisement in the advertisement placement playlist according to the matching degree score of each candidate advertisement includes: The reference display duration of each candidate advertisement is determined according to the total advertisement duration constraint and the number of candidate advertisements; The reference display duration is dynamically adjusted according to the matching degree score of each candidate advertisement to obtain the actual display duration of each candidate advertisement.

[0014] In a possible implementation, the first advertisement placement strategy is further optimized according to the user gaze duration data to obtain a second advertisement placement strategy, including: The placement efficiency value of each candidate advertisement is determined according to the user gaze duration data and the actual display duration of each candidate advertisement, where the placement efficiency value is the ratio of the average user gaze duration to the actual display duration; The first advertisement placement strategy is further optimized according to the placement efficiency value of each candidate advertisement to obtain a second advertisement placement strategy.

[0015] In a possible implementation, the first advertisement placement strategy is further optimized according to the placement efficiency value of each candidate advertisement to obtain a second advertisement placement strategy, including: When the placement efficiency value of the candidate advertisement exceeds 1.2 times the average value of the same type of advertisement in N consecutive monitoring periods, a progressive display duration gain strategy is triggered; When the delivery efficiency value of the candidate advertisement is lower than 0.8 times the average value of the same type of advertisement within M consecutive monitoring cycles, the display duration grading attenuation strategy is triggered, where 5≥N≥M≥3.

[0016] In a second aspect, an advertisement delivery strategy optimization system is further provided in an embodiment of the present application, including: a memory and a processor, where the memory is used to store program code; the processor is used to call the program code to execute the method described in the first aspect.

[0017] Different from the prior art, the advertisement delivery strategy optimization method based on AI learning provided in the embodiment of the present application first uses multi-source environmental sensors to capture real-time state parameters of the external environment, combines big data analysis to associate with hot online shopping data, and then quantifies the matching degree between each candidate advertisement and the environmental state through a pre-trained model to generate a first delivery strategy; subsequently, during the advertisement display process, taking implicit behavior data such as user gaze duration as a feedback index, the first delivery strategy is secondarily optimized to obtain a second delivery strategy. In this way, by combining environmental parameters and online shopping hot data, and using model scoring and ranking to optimize the initial strategy, the advertisement delivery is more in line with the environmental state and online shopping consumption trend, greatly improving the adaptability of advertisement delivery to the scenario environment and online shopping hotspots; and it can accurately capture user interests, enhance the attractiveness and pertinence of advertisements, thereby improving the advertisement delivery effect and user attention, and realizing the intelligent and precise optimization of the advertisement delivery strategy. Description of the Drawings

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on the structures shown in these drawings without creative efforts.

[0019] Figure 1 It is a schematic flowchart of the advertisement delivery strategy optimization method based on AI learning in some embodiments of the present application; Figure 2 It is a schematic flowchart of step S100 of the advertisement delivery strategy optimization method based on AI learning in some embodiments of the present application; Figure 3 It is a schematic flowchart of step S300 of the advertisement delivery strategy optimization method based on AI learning in some other embodiments of the present application; Figure 4 It is a schematic flowchart of step S400 of the advertisement delivery strategy optimization method based on AI learning in some other embodiments of the present application; Figure 5This is a schematic diagram of the hardware structure of the advertising placement strategy optimization system in some embodiments of the present application.

[0020] The implementation, functional features, and advantages of the present invention will be further described with reference to the embodiments and the accompanying drawings. Specific embodiments

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0022] It should be noted that all directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention are only used to explain the relative positional relationship and movement conditions between components in a specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0023] In addition, the descriptions involving "first", "second", etc. in the present invention are only for descriptive purposes, and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In addition, "and / or" throughout the text includes three scenarios. Taking A and / or B as an example, it includes the technical solution of A, the technical solution of B, and the technical solution that A and B are satisfied at the same time. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0024] In the digital marketing era, the accuracy and efficiency of advertising placement are directly related to the operational benefits of business entities. Traditional shopping mall advertising placement strategies usually arrange content based on fixed time periods, preset rules, or manual experience, and it is difficult to dynamically adapt to environmental changes and real-time fluctuations in user behavior. For example, seasonal changes, weather changes, or emergencies may cause significant changes in consumers' shopping preferences, while static advertising strategies cannot capture such environmental correlations in a timely manner, resulting in waste of resources and low conversion rates. At the same time, the booming development of online consumer behavior has made users' shopping decisions show the characteristics of "online-offline integration", but offline advertising systems lack the ability to integrate real-time online shopping trends, resulting in insufficient matching between advertising content and users' potential needs.

[0025] In the operation and management of a shopping mall, the environmental status has a non-negligible impact on product sales. Different environmental statuses (such as indoor, outdoor environments or changes in the corresponding environments) will directly or indirectly affect consumers' behaviors and demands, and thus affect the sales performance of products. In order to achieve precise optimization of advertising placement and improve the operation efficiency and sales performance of the shopping mall, this application takes obtaining the environmental status parameters collected in real time by environmental sensors as the starting point to provide data support for the formulation of subsequent advertising strategies.

[0026] Such as Figures 1-4 shown, the following takes the advertising placement strategy optimization system executing the advertising placement strategy optimization method based on AI learning as an example for illustration. It should be noted that although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from here. Please refer to the appendix Figure 1 and this method includes the following steps S100 - step S600: Step S100, obtain the environmental status parameters collected in real time by environmental sensors, where the environmental status parameters include at least one of indoor environmental parameters and outdoor environmental parameters; It can be understood that the environmental statuses inside or outside the shopping mall are different, or rather, the environmental statuses inside or outside the shopping mall change, which have different impacts on the product sales in the shopping mall. For example, in hot weather with high outdoor temperatures, if the shopping mall can prioritize advertising products related to "coping with the heat" in advertising placement, such as cold drinks, sunscreen products, portable small fans, etc., it will better meet the actual needs of consumers at present, thus effectively attracting customers' attention, stimulating the desire to purchase, and then significantly improving the overall operation efficiency and sales performance of the shopping mall.

[0027] Based on this, the embodiments of this application first obtain the environmental status parameters collected in real time by environmental sensors to optimize and adjust advertising placement according to the environmental status parameters.

[0028] An environmental sensor refers to a sensor used to monitor the environmental status, which can be a people flow detection sensor for monitoring the indoor people flow density, a temperature sensor for detecting the indoor or outdoor environmental temperature, or a rainfall detection sensor for detecting the outdoor rainfall.

[0029] The people flow detection sensor is mainly used to monitor the indoor people flow density. For example, the people flow density can be detected through image analysis. By real-time sensing the personnel flow conditions in different areas of the shopping mall, it can accurately reflect the popularity of each area and the gathering situation of consumers.

[0030] In one embodiment, the environmental sensor includes an indoor environmental sensor and an outdoor environmental sensor. Step S100: Obtain the environmental status parameters collected in real time by environmental sensors, including: S110. Obtain the indoor environmental status parameters collected in real time by the indoor environmental sensors, where the indoor environmental status parameters at least include the indoor population density; S120. Obtain the outdoor environmental status parameters collected in real time by the outdoor environmental sensors, where the outdoor environmental status parameters at least include one of the environmental temperature and the environmental rainfall.

[0031] Specifically, indoor environmental sensors (such as population detection sensors) can be distributed in various key areas of the shopping mall, such as entrances, elevator entrances, main passages, and around popular stores on each floor. These sensors continuously monitor the indoor environment and transmit the collected population density data to the system processor in real time, thereby realizing the detection of the population density in the shopping mall.

[0032] Outdoor environmental sensors (such as temperature sensors and rainfall detection sensors) are installed in suitable positions outside the shopping mall and can accurately sense the outdoor environmental changes. The temperature sensor monitors the outdoor temperature in real time and transmits the data to the system processor; the rainfall detection sensor automatically starts when it rains, records the rainfall information, and uploads the relevant data in a timely manner.

[0033] In the embodiment of the present application, a population detection sensor is set indoors to detect the population density in the shopping mall indoors, and a temperature sensor or a rainfall sensor is set outdoors to detect the temperature or rainfall data outside the shopping mall, providing a solid and reliable data support and decision-making basis for formulating a more targeted advertising placement strategy that better meets the needs of consumers in the future.

[0034] Exemplarily, in a certain area where the shopping mall is located, there are many heavy rains in summer. When the rainfall detection sensor detects the approaching heavy rain, the shopping mall can quickly display various preferential activities of rain gear through the billboards at the shopping mall entrance; at the same time, promote the exclusive entertainment packages for rainy days in areas such as cinemas and game halls in the shopping mall, guiding consumers to enjoy the fun of shopping and entertainment while taking shelter from the rain.

[0035] Step S200. Obtain the hot online shopping data that matches the environmental status parameters according to the environmental status parameters; After obtaining the indoor and outdoor environmental status parameters of the shopping mall, in order to further improve the overall operation efficiency and sales performance of the shopping mall, the embodiment of the present application further obtains the hot online shopping data that matches these environmental status parameters based on these environmental status parameters. By combining the environmental status with the online consumption trend, it can more accurately insight into the needs of consumers, thereby providing strong support for the subsequent advertising placement strategy.

[0036] In one embodiment, obtaining the hot online shopping data that matches the environmental status parameters according to the environmental status parameters includes: When the indoor population density is greater than or equal to the population density threshold and the environmental temperature is greater than or equal to the temperature threshold, a reinforcement learning search algorithm is used to obtain the data of hot-selling products in online shopping during high-temperature weather; Alternatively, when the indoor population density is greater than or equal to the population density threshold and the environmental rainfall is greater than or equal to the rainfall threshold, a reinforcement learning search algorithm is used to obtain the data of hot-selling products in online shopping during rainy weather.

[0037] In an actual operation scenario, when the indoor population density reaches or exceeds the preset population density threshold and the outdoor environmental temperature reaches or exceeds the preset temperature threshold, it indicates that there is a large number of people in the mall and the outdoor environmental temperature is high. In this case, the shopping needs and consumption preferences of consumers are very likely to change significantly due to the high temperature. To accurately capture this change, the embodiment of this application introduces a reinforcement learning search algorithm to deeply explore the massive data of major online shopping platforms and mine the data of hot-selling products in high-temperature weather. Taking a sweltering summer with continuous high temperature as an example, the high temperature will prompt consumers to pay more attention to products for relieving heat and cooling down. Cooling drinks, sunscreen, lightweight and portable small fans, etc. become popular choices on their shopping lists. Through the reinforcement learning search algorithm, these hot-selling products can be efficiently and accurately locked, providing a strong basis for subsequent marketing decisions.

[0038] In another actual operation scenario, when the indoor population density reaches or exceeds the preset threshold and the outdoor environmental rainfall reaches or exceeds the preset rainfall threshold, it means that there are many consumers in the mall while it is rainy outdoors. Rainfall will not only affect consumers' travel plans but also change their shopping needs. At this time, the reinforcement learning search algorithm is also used to deeply scan and search the online shopping platform to obtain the data of hot-selling products in rainy weather. For example, in rainy days, rain gear such as umbrellas, raincoats, and waterproof shoe covers becomes essential, and consumers' demand for them will increase significantly; at the same time, due to inconvenient travel, people tend to engage in indoor entertainment activities, so indoor entertainment devices such as game consoles and books may also become hot-selling products.

[0039] In other embodiments, if the indoor population density does not exceed the preset population density threshold, it means that there are few people in the mall, and the optimization of the advertising placement strategy can be not carried out, that is, the initial advertising placement strategy can be maintained for advertising placement.

[0040] The reinforcement learning search algorithm in the embodiment of this application can efficiently screen out hot-selling products with a high degree of matching with environmental state parameters from massive online shopping data by continuously learning and optimizing the search strategy. This algorithm can combine historical sales data, user behavior data, and real-time environmental parameters to dynamically adjust the search weights, ensuring that the obtained data is both timely and highly relevant.

[0041] It should be noted that the hot online shopping data may include multiple product names that match the environmental state parameters and data such as the popularity of the corresponding products.

[0042] Exemplarily, the following is a data table of the best-selling products in online shopping during high-temperature weather obtained by the reinforcement learning search algorithm.

[0043]

[0044] Step S300: Input the environmental state parameters and the hot online shopping data into a pre-trained advertisement matching model, and output the matching degree scores of each candidate advertisement. After obtaining the hot online shopping data that matches the environmental state parameters, the embodiment of the present application inputs the environmental state parameters and the hot online shopping data into a pre-trained advertisement matching model, and outputs the matching degree scores of each candidate advertisement. In this way, by fusing the environmental state parameters and the hot online shopping data, the comprehensive matching degree scores of each candidate advertisement are calculated, providing a basis for subsequent advertisement placement optimization.

[0045] In one embodiment, step S300: Input the environmental state parameters and the hot online shopping data into a pre-trained advertisement matching model, and output the matching degree scores of each candidate advertisement, including: S310: Input the environmental state parameters into the environmental sub-network of the pre-trained advertisement matching model to obtain the environmental attribute scores of each candidate advertisement. S320: Input the hot online shopping data into the hot sub-network of the pre-trained advertisement matching model to obtain the hot attribute scores of each candidate advertisement. S330: Perform weighted summation on the environmental attribute scores and the hot attribute scores to obtain the matching degree scores of each candidate advertisement.

[0046] Specifically, in the embodiment of the present application, the environmental state parameters are first input into the environmental sub-network of the pre-trained advertisement matching model to obtain the environmental attribute scores of each candidate advertisement. Then, the hot online shopping data is input into the hot sub-network of the pre-trained advertisement matching model to obtain the hot attribute scores of each candidate advertisement. Finally, weighted summation is performed on the environmental attribute scores and the hot attribute scores to obtain the matching degree scores of each candidate advertisement.

[0047] Furthermore, if a shopping mall has 10 advertisements to be placed, they are restaurant A (providing cold drinks), restaurant B (providing tea drinks), restaurant C (providing normal temperature water), daily necessities store A (selling sunscreen clothing, etc.), daily necessities store B (selling small fans, umbrellas, etc.), clothing store A, clothing store B, skin care store A (specializing in selling sunscreen), skin care store B (selling a small amount of sunscreen and other skin care products), and beverage store A. If the environmental state parameters detected by the current environmental sensor indicate that the outside world is hot weather, and the matched hot online shopping data indicates that the current online shopping hot-selling products are ranked according to the popularity of portable fans, sunscreens, and sun-proof clothing. Inputting the environmental state parameters into the pre-trained advertising matching model environment subnetwork to obtain the environmental attribute scores of each candidate advertisement means that the environmental attribute scores of each candidate advertisement are obtained by model matching according to the environmental state parameters. The higher the degree of matching between the environmental state parameters and the candidate advertisement, the higher its score value. For example, when the ambient temperature is high, advertisements corresponding to stores related to summer heat relief (such as stores selling small fans, sunscreen, sun umbrellas, and sun-protective clothing) have a higher matching degree, while advertisements unrelated to high temperatures (such as room temperature drinks and ordinary clothing) have a lower matching degree.

[0048] Exemplarily, the scores of the above 10 candidate advertisements are obtained by model matching according to the environmental state parameters, that is, the environmental attribute scores are: daily necessities store A gets 100 points, daily necessities store B gets 100 points, skin care store A gets 100 points, beverage store A gets 100 points, skin care store B gets 80 points, restaurant A gets 60 points, restaurant C gets 40 points, restaurant B gets 20 points, clothing store A gets 0 points, and clothing store B gets 0 points.

[0049] By matching the model with the hot online shopping data, we can obtain the scores of the above 10 candidate advertisements, that is, the hot attribute scores are: daily necessities store A got 80 points, daily necessities store B got 100 points, skin care store A got 90 points, skin care store B got 90 points, restaurant A got 50 points, restaurant C got 40 points, restaurant B got 30 points, clothing store A got 0 points, and clothing store B got 0 points.

[0050] After obtaining the environmental attribute score and hotspot attribute score of each candidate advertisement, a weighted sum can be performed according to preset weights (such as 0.5 for each weight) to obtain the matching score of each candidate advertisement.

[0051] In the embodiment of the present application, each candidate advertisement is comprehensively scored according to the environmental state and the hot online shopping data matching the environmental state to obtain a matching score for each candidate advertisement, thereby providing accurate data support for the optimization of subsequent advertising delivery strategies.

[0052] In one embodiment, before obtaining the matching degree scores of each candidate advertisement by performing weighted summation on the environmental attribute scores and the hot spot attribute scores, it further includes: obtaining the change rate of the environmental state parameters and the trend acceleration of the hot spot online shopping data; inputting the change rate of the environmental state parameters and the trend acceleration of the hot spot online shopping data into a dynamic selection network to generate dynamic weight coefficients, where the dynamic selection network satisfies the following expression: ;

[0053] ;

[0054] ;

[0055] In the formula, is the dynamic weight of the environmental attribute score, is the dynamic weight of the hot spot attribute score; is the change rate of the environmental state parameters, is the reference change rate of the environmental state parameters, is the reference weight of the environmental attribute score; is the trend acceleration of the hot spot online shopping data, is the reference trend acceleration of the hot spot online shopping data, is the reference weight of the hot spot attribute score; is the change rate warning value of the environmental state parameters, is the trend acceleration warning value of the hot spot online shopping data.

[0056] Specifically, when the change rate (Φ) of the environmental state parameters increases significantly (exceeds Φ1), and the hot spot trend acceleration H is less than H1, such as when the change rate of the environmental temperature increases significantly but the hot spot trend is not obvious, the weight value α corresponding to the environmental attribute score increases, that is, the influence of the environmental attribute score is enhanced; at this time, the dynamic weight of the environmental attribute score is calculated according to the reference weight of the environmental attribute score. When the hot spot trend acceleration (H) suddenly increases (exceeds H1), and the change rate Φ of the environmental state parameters is less than Φ1, such as when the sales growth rate of online shopping products matching the environment increases significantly, the weight value β corresponding to the hot spot attribute score increases, that is, the influence of the hot spot attribute score is enhanced. At this time, the dynamic weight of the hot spot attribute score is calculated according to the reference weight of the hot spot attribute score.

[0057] In this way, the advertisement placement strategy optimization system of the present application can quickly adapt to sudden changes in the external environment and online shopping platforms (such as sudden weather changes, holiday promotions, etc.), and avoid score rigidity caused by fixed weights. And the real-time data-driven decision-making in the embodiments of the present application ensures that the advertisement matching strategy is always synchronized with the current environment and the dynamics of the online shopping platform, thereby further improving the accuracy of the advertisement placement strategy optimization.

[0058] Step S400: Sort the candidate ads according to the matching degree scores, and generate a first ad placement strategy to preliminarily optimize the initial ad placement strategy; After obtaining the matching degree scores of each candidate ad with the current environmental state and the corresponding hot online shopping data, the initial optimization of the ad strategy placement can be performed according to the matching degree scores, such as adjusting the ad placement order and placement duration, etc.

[0059] In one embodiment, step S400: Sort the candidate ads according to the matching degree scores, and generate a first ad placement strategy to preliminarily optimize the initial ad placement strategy, including: S410: Sort the candidate ads according to the descending order of the matching degree scores to obtain an ad placement playlist; S420: Determine the display duration of each candidate ad in the ad placement playlist according to the matching degree score of each candidate ad.

[0060] Specifically, the candidate ads can be sorted according to the descending order of the matching degree scores to obtain an ad placement playlist first. The candidate ads with high matching degree scores are arranged in the front to be preferentially displayed, and the candidate ads with low matching degree scores are arranged in the back. Then, determine the display duration of each candidate ad in the ad placement playlist according to the matching degree score of each candidate ad. The higher the matching degree score of the candidate ad, the higher its matching degree with the environmental state, and the longer its display duration. In this way, by adjusting the candidate ads with high matching degree with the environmental state to be preferentially displayed and increasing the display duration, the exposure duration of such ads is increased, thereby improving the sales conversion of the corresponding products.

[0061] When determining the display duration of each candidate ad in the ad placement playlist according to the matching degree score of each candidate ad, the benchmark display duration of each candidate ad can be determined first according to the total ad duration constraint and the number of candidate ads; then, the benchmark display duration is dynamically adjusted according to the matching degree score of each candidate ad to obtain the actual display duration of each candidate ad.

[0062] Specifically, within a total ad duration constraint period, a benchmark display duration is evenly allocated to the candidate ads to be displayed. Then, the benchmark display duration is dynamically adjusted according to the matching degree score of each candidate ad to obtain the actual display duration of each candidate ad. Among them, the display duration can be reallocated according to the ratio between the matching degree scores of the candidate ads.

[0063] Exemplarily, under the time constraint of cycling through 10 candidate ads within 10 minutes, the benchmark display duration is 1 minute for each candidate ad. In the embodiments of the present application, the display duration is reallocated according to the ratio between the matching degree scores of each candidate ad, and the specific process will not be elaborated.

[0064] In this way, the display order of the adjusted candidate ads and the corresponding actual display duration are the first advertising placement strategy. Therefore, subsequently, the playback content of the advertising display screen can be controlled according to this first advertising placement strategy, that is, advertising is placed on the advertising display screen according to this first advertising placement strategy.

[0065] Step S500: Control the playback content of the advertising display screen according to the first advertising placement strategy, and collect user gaze duration data in real time during the playback process; After advertising is placed on the advertising display screen according to the first advertising placement strategy, the user gaze duration data for each candidate ad is collected in real time. The gaze duration data can be the cumulative sum of all users' gaze durations within one display cycle, or the average value of all users' gaze durations within one display cycle, or the average value of all users' gaze durations within multiple consecutive display cycles.

[0066] Step S600: Optimize the first advertising placement strategy again according to the user gaze duration data to obtain a second advertising placement strategy.

[0067] Specifically, the advertising placement efficiency value of each candidate ad can be determined first according to the user gaze duration data and the actual display duration of each candidate ad. Among them, the advertising placement efficiency value is the ratio of the average user gaze duration to the actual display duration. Then, the first advertising placement strategy is optimized again according to the advertising placement efficiency values of each candidate ad to obtain a second advertising placement strategy. For example, when the advertising placement efficiency value of a candidate ad exceeds 1.2 times the average value of similar ads in N consecutive monitoring cycles, a progressive display duration gain strategy is triggered; when the advertising placement efficiency value of a candidate ad is lower than 0.8 times the average value of similar ads in M consecutive monitoring cycles, a display duration hierarchical decay strategy is triggered, where 5 ≥ N ≥ M ≥ 3.

[0068] Exemplarily, if the average user gaze duration of candidate advertisement A within one delivery cycle is 1 minute and its actual display duration is 3 minutes, then its delivery efficiency value is 0.33. The calculation of the delivery efficiency values of other candidate advertisements is similar and will not be elaborated here. If the delivery efficiency value of candidate advertisement A exceeds 1.2 times the average value of similar advertisements in 4 consecutive display cycles, it indicates that the input-output ratio of this advertisement is relatively high, and then the progressive display duration gain strategy is triggered. For example, in subsequent advertisement cycles, the display duration is increased according to the display duration increase strategies of 10% and 20%. If the delivery efficiency value of candidate advertisement A is lower than 0.8 times the average value of similar advertisements in 3 consecutive display cycles, it indicates that the input-output ratio of this advertisement is relatively low, and then the display duration grading decay strategy is triggered. For example, in subsequent advertisement cycles, the display duration is decreased according to the display duration decay strategies of 10% and 20%. In this way, advertisements that users are more interested in obtain more display opportunities, while advertisements that users are not interested in obtain fewer display opportunities, thereby achieving the secondary optimization of the advertisement delivery strategy.

[0069] Based on this, in the embodiments of the present application, the initial advertisement delivery strategy is initially optimized according to the environmental state parameters and the matching hot online shopping data to obtain the first advertisement delivery strategy. Then, the second advertisement delivery strategy is obtained by further optimizing according to the delivery efficiency values of each candidate advertisement. In this way, the advertisement delivery effect and user attention are greatly improved, and the intelligent and precise optimization of the advertisement delivery strategy is achieved.

[0070] As Figure 5 shown, Figure 5 is a schematic diagram of the hardware structure of the advertisement delivery strategy optimization system in some embodiments of the present application. The advertisement delivery strategy optimization system provided in the embodiments of the present application further includes a memory 1000 and a processor 2000. Among them, the memory 1000 is used to store computer-readable instructions, and the processor 2000 is used to call the computer-readable instructions to execute the advertisement delivery strategy optimization method based on AI learning as described above.

[0071] Among them, the processor 2000 is used to provide computing and control capabilities to control the advertisement placement strategy optimization system to execute corresponding tasks. For example, it controls the advertisement placement strategy optimization system to execute the advertisement placement strategy optimization method based on AI learning in any of the above method embodiments. The method includes: obtaining environmental state parameters of the target area collected in real time by an environmental sensor, where the environmental state parameters include at least one of indoor environmental parameters and outdoor environmental parameters; obtaining hot online shopping data matching the environmental state parameters according to the environmental state parameters; inputting the environmental state parameters and the hot online shopping data into a pre-trained advertisement matching model to output the matching degree scores of each candidate advertisement; sorting the priorities of each candidate advertisement according to the matching degree scores to generate a first advertisement placement strategy to preliminarily optimize the initial advertisement placement strategy; controlling the playing content of the advertisement placement screen according to the first advertisement placement strategy, and collecting user gaze duration data in real time during the playing process; and re-optimizing the first advertisement placement strategy through the user gaze duration data to obtain a second advertisement placement strategy.

[0072] The processor 2000 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a hardware chip, or any combination thereof; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The above PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0073] The memory 1000, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs, non-transitory computer-executable programs, and modules, such as the program instructions / modules corresponding to the advertisement placement strategy optimization method based on AI learning in the embodiments of the present application. By running the non-transitory software programs, instructions, and modules stored in the memory 1000, the processor 2000 can implement the advertisement placement strategy optimization method based on AI learning in any of the above method embodiments.

[0074] Specifically, the memory 1000 may include a volatile memory (VM), such as a random access memory (RAM); the memory 1000 may also include a non-volatile memory (NVM), such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid-state drive (SSD), or other non-transitory solid-state storage devices; the memory 1000 may further include a combination of the above types of memories.

[0075] In summary, the advertising placement strategy optimization system of the present application adopts the technical solution of any one of the above embodiments of the advertising placement strategy optimization method based on AI learning. Therefore, it has at least the beneficial effects brought by the technical solutions of the above embodiments, which will not be elaborated herein one by one.

[0076] The embodiment of the present application also provides a computer-readable storage medium, such as a memory including program code, and the above program code can be executed by a processor to complete the advertising placement strategy optimization method based on AI learning in the above embodiment. For example, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CDROM), a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0077] The embodiment of the present application also provides a computer program product, which includes one or more pieces of program code, and the program code is stored in a computer-readable storage medium. The processor of the warning system reads the program code from the computer-readable storage medium, and the processor executes the program code to complete the steps of the advertising placement strategy optimization method provided in the above embodiment.

[0078] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above embodiments can be completed by hardware, or can be completed by hardware related to program code. The program can be stored in a computer-readable storage medium, and the above-mentioned storage medium can be a read-only memory, a magnetic disk, or an optical disc, etc.

[0079] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0080] Through the description of the above embodiments, those of ordinary skill in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and of course, it can also be implemented by hardware. Those of ordinary skill in the art can understand that all or part of the processes of implementing the above method embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above method embodiments. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0081] The above is only the preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structural transformation made by using the description of the present invention and the content of the drawings under the inventive concept of the present invention, or direct / indirect application in other related technical fields, is included in the patent protection scope of the present invention.

Claims

1. An optimization method for advertising placement strategies based on AI learning, which is applied to an advertising placement strategy optimization system. The advertising placement strategy optimization system includes an environmental sensor and an advertising display screen, and is characterized in that, The method includes: Obtaining environmental status parameters collected in real time by an environmental sensor, where the environmental status parameters include at least one of indoor environmental parameters and outdoor environmental parameters; Obtaining hot online shopping data that matches the environmental status parameters according to the environmental status parameters; Inputting the environmental status parameters and the hot online shopping data into a pre-trained advertisement matching model to output the matching degree scores of each candidate advertisement; Performing priority ranking on each candidate advertisement according to the matching degree scores to generate a first advertisement placement strategy for initially optimizing the initial advertisement placement strategy; Controlling the playback content of the advertisement placement screen according to the first advertisement placement strategy, and collecting user gaze duration data in real time during the playback process; Performing re-optimization on the first advertisement placement strategy according to the user gaze duration data to obtain a second advertisement placement strategy.

2. The method for optimizing an advertising placement strategy based on AI learning according to claim 1, characterized in that, The environmental sensor includes an indoor environmental sensor and an outdoor environmental sensor. Obtaining the environmental status parameters collected in real time by the environmental sensor includes: Obtaining the indoor environmental status parameters collected in real time by the indoor environmental sensor, where the indoor environmental status parameters at least include the indoor population density; Obtaining the outdoor environmental status parameters collected in real time by the outdoor environmental sensor, where the outdoor environmental status parameters at least include one of the environmental temperature and the environmental rainfall.

3. The method for optimizing an advertising placement strategy based on AI learning according to claim 2, wherein Obtaining hot online shopping data that matches the environmental status parameters according to the environmental status parameters includes: When the indoor population density is greater than or equal to the population density threshold and the environmental temperature is greater than or equal to the temperature threshold, using a reinforcement learning search algorithm to obtain the hot-selling product data of online shopping in high-temperature weather; Or, when the indoor population density is greater than or equal to the population density threshold and the environmental rainfall is greater than or equal to the rainfall threshold, using a reinforcement learning search algorithm to obtain the hot-selling product data of online shopping in rainy weather.

4. The method for optimizing an advertising placement strategy based on AI learning according to claim 1, characterized in that, Inputting the environmental status parameters and the hot online shopping data into a pre-trained advertisement matching model to output the matching degree scores of each candidate advertisement includes: Inputting the environmental status parameters into the environmental sub-network of the pre-trained advertisement matching model to obtain the environmental attribute scores of each candidate advertisement; Inputting the hot online shopping data into the hot sub-network of the pre-trained advertisement matching model to obtain the hot attribute scores of each candidate advertisement; Performing weighted summation on the environmental attribute scores and the hot attribute scores to obtain the matching degree scores of each candidate advertisement.

5. The method for optimizing an advertising placement strategy based on AI learning according to claim 4, wherein Before performing weighted summation on the environmental attribute scores and the hot attribute scores to obtain the matching degree scores of each candidate advertisement, it further includes: Obtaining the change rate of the environmental status parameters and the trend acceleration of the hot online shopping data; Inputting the change rate of the environmental status parameters and the trend acceleration of the hot online shopping data into a dynamic selection network to generate dynamic weight coefficients, where the dynamic selection network satisfies the following expression: ； ； ； In the formula, is the dynamic weight of the environmental attribute score, is the dynamic weight of the hot spot attribute score; is the change rate of the environmental state parameter, is the reference change rate of the environmental state parameter, is the reference weight of the environmental attribute score; is the trend acceleration of the hot spot online shopping data, is the reference trend acceleration of the hot spot online shopping data, is the reference weight of the hot spot attribute score; is the warning value of the change rate of the environmental state parameter, is the warning value of the trend acceleration of the hot spot online shopping data.

6. The method for optimizing an advertising placement strategy based on AI learning according to claim 1, characterized in that, Performing priority ranking on candidate advertisements according to the matching degree scores to generate a first advertisement placement strategy for initially optimizing the initial advertisement placement strategy includes: Performing priority ranking on candidate advertisements in descending order of the matching degree scores to obtain an advertisement placement playback list; Determine the display duration of each candidate advertisement in the advertisement placement playlist according to the matching degree score of each candidate advertisement.

7. The method for optimizing an advertising placement strategy based on AI learning according to claim 6, characterized in that, Determine the display duration of each candidate advertisement in the advertisement placement playlist according to the matching degree score of each candidate advertisement, including: Determine the benchmark display duration of each candidate advertisement according to the total advertisement duration constraint and the number of candidate advertisements; Dynamically adjust the benchmark display duration according to the matching degree score of each candidate advertisement to obtain the actual display duration of each candidate advertisement.

8. The method for optimizing an advertising placement strategy based on AI learning according to claim 7, wherein Re-optimize the first advertisement placement strategy according to the user gaze duration data to obtain a second advertisement placement strategy, including: Determine the placement efficiency value of each candidate advertisement according to the user gaze duration data and the actual display duration of each candidate advertisement, where the placement efficiency value is the ratio of the average user gaze duration to the actual display duration; Re-optimize the first advertisement placement strategy according to the placement efficiency value of each candidate advertisement to obtain a second advertisement placement strategy.

9. The method for optimizing an advertising placement strategy based on AI learning according to claim 8, wherein Re-optimize the first advertisement placement strategy according to the placement efficiency value of each candidate advertisement to obtain a second advertisement placement strategy, including: When the placement efficiency value of a candidate advertisement exceeds 1.2 times the average value of similar advertisements in N consecutive monitoring cycles, trigger a progressive display duration gain strategy; When the placement efficiency value of a candidate advertisement is lower than 0.8 times the average value of similar advertisements in M consecutive monitoring cycles, trigger a display duration hierarchical decay strategy, where 5 ≥ N ≥ M ≥ 3.

10. An advertising placement strategy optimization system, characterized in that, Including: A memory and a processor, the memory is used to store program code; The processor is used to call the program code to execute the method according to any one of claims 1 to 9.

Images