An advertisement dynamic delivery method and system based on environment perception

Abstract

The application discloses an advertisement dynamic putting method and system based on environment perception, relates to the field of advertisement putting, and solves the technical problem that it is difficult to improve advertisement pushing accuracy according to surrounding environment perception ability in advertisement putting; the application plans a putting interconnection area, identifies whether there is an environment factor guiding a user in the putting interconnection area; if yes, identifies whether the guiding effect of the environment factor is a long-time tendency or a short-time tendency; when it is the long-time tendency, related advertisements are put to the user entering the putting interconnection area; when it is the short-time tendency, a stop time of the short-time tendency is predicted based on user behavior data, related advertisements are put to the user entering the putting interconnection area before the stop time, and related advertisements are dynamically put to the user entering the putting interconnection area after the stop time; if not, advertisements are put according to preset advertisements; and the application can improve advertisement pushing accuracy.

Description

Technical Field

[0001] This invention relates to the field of advertising delivery, specifically a method and system for dynamic advertising delivery based on environmental perception. Background Technology

[0002] The digital advertising industry has developed rapidly with the popularization of the Internet and mobile devices. However, traditional advertising methods often lack targeting and real-time capabilities, resulting in poor advertising effects and wasted resources. At the same time, with the increasing demand for personalized experiences from consumers and the intensification of market competition, advertisers urgently need smarter and more precise advertising solutions. Meanwhile, most existing advertising systems are based on users' historical behavior and demographic characteristics for targeting, lacking consideration of real-time environmental factors, and thus failing to meet advertisers' needs for precise targeting and maximum effectiveness.

[0003] Currently, most environmentally-aware dynamic ad delivery methods struggle to improve ad delivery accuracy by leveraging the ability to perceive the surrounding environment. More specifically, they fail to identify environmental factors influencing user preferences based on the advertising inclinations of surrounding people, and thus cannot intelligently push ads. This results in a lack of intelligent matching between ad content and the user's environment, reducing ad relevance, user experience, and overall campaign effectiveness.

[0004] Therefore, this invention discloses a dynamic advertising delivery method and system based on environmental perception to solve the above-mentioned technical problems. Summary of the Invention

[0005] The present invention aims to at least solve one of the technical problems existing in the prior art; to this end, the present invention proposes an environmental perception-based dynamic advertising delivery method and system to solve the technical problem that it is difficult to improve the accuracy of advertising delivery based on the ability to perceive the surrounding environment.

[0006] To achieve the above objectives, a first aspect of the present invention provides a method for dynamic advertising delivery based on environmental awareness, comprising: Plan the interconnected advertising areas and acquire user behavior data within those areas; the user behavior data includes the number of user searches for each ad, the number of user clicks on each ad, and the time users spend on the ads. Identify whether there are environmental factors guiding users within the targeted interconnection area based on the historical user search counts of users within the targeted interconnection area; Yes, it identifies whether the guiding effect of environmental factors is a long-term or short-term tendency; when it is a long-term tendency, it delivers relevant advertisements to users entering the advertising interconnection area; when it is a short-term tendency, it predicts the stopping time of the short-term tendency based on user behavior data, delivers relevant advertisements to users entering the advertising interconnection area before the stopping time, and dynamically delivers relevant advertisements to users entering the advertising interconnection area after the stopping time. No, the ads will be placed according to the preset ads.

[0007] Preferably, the planning of the interconnection area includes: Obtain geographic map data within the target delivery area, and divide the target delivery area into multiple geographic grids of a preset size; wherein the target delivery area is defined manually or generated based on administrative division data; Based on the geographic map data, the geographical locations and types of various infrastructures within the target deployment area are identified; the types include stadiums, shopping malls, schools, hospitals, office buildings, and communities. Based on the type of infrastructure, the corresponding influence radius parameter is invoked; taking the geographical location of the infrastructure as the center, the coverage area of ​​the corresponding infrastructure is determined according to the influence radius parameter; wherein, the influence radius parameter is determined based on the historical pedestrian density of infrastructure of the same type; Traverse all geographic grids and identify whether a geographic grid falls within the coverage of one or more infrastructures. If a geographic grid falls within the coverage of only one infrastructure, mark that geographic grid as the target grid of that infrastructure. If a geographic grid falls within the coverage of multiple infrastructures, mark the infrastructure closest to the current geographic grid as the target grid of the current grid. Determine the deployment interconnection area of ​​the infrastructure based on the target grids of each infrastructure.

[0008] Preferably, the acquisition of user behavior data within the interconnected area includes: After user authorization, when the location device identifies the user entering the advertising interconnection area, the SDK data points and server logs are used to obtain the number of user searches for each advertisement, the number of user clicks on each advertisement, and the time the user spends in the advertisement within the advertising interconnection area.

[0009] Preferably, identifying whether there are environmental factors guiding users within the advertising interconnection area based on the historical user search frequency of users within the advertising interconnection area includes: The system collects historical user search counts for each advertisement within the previous historical monitoring window in the target area. Advertisements are extracted sequentially and marked as target advertisements. The historical user search counts for these target advertisements are aggregated according to time granularity to construct a search popularity time series for the target advertisements. Statistical characteristics of each search popularity time series are obtained and marked as statistical characteristic one. The statistical characteristics include the mean and maximum value of user search counts, forming an initial vector describing the search behavior characteristics of the current area. The historical monitoring window is manually set, ranging from 3 to 10 days. The time granularity is manually set, and can be 1 hour or 30 minutes. The statistical characteristic is the historical user search counts included in the target advertisement within the corresponding search popularity time series. The system acquires statistical characteristics of the search popularity time sequence of the current advertising region within the n historical monitoring windows prior to the current time, and marks these statistical characteristics as statistical characteristic two. When the deviation factor of the statistical characteristics in statistical characteristic one and statistical characteristic two is greater than the deviation threshold, it is determined that there are environmental factors in the current advertising region that guide users towards the target advertisement; when the deviation factor of the statistical characteristics in statistical characteristic one and statistical characteristic two is not greater than the deviation threshold, it is determined that there are no environmental factors in the current advertising region that guide users towards the target advertisement. Here, n is manually set and can take values ​​ranging from [value missing]. 3. The deviation threshold is determined based on the statistical distribution of the deviation factor calculated historically. The formula for calculating the deviation factor is: PL=(JZ1-JZ2) / BJZ+(DZ1-DZ2) / BDZ. PL is the deviation factor, JZ1 is the mean in statistical feature one, JZ2 is the mean in statistical feature two, DZ1 is the maximum value in statistical feature one, DZ2 is the maximum value in statistical feature two, BJZ is the standard mean difference obtained from historical data in the current deployment interconnection area, and BDZ is the standard maximum difference obtained from historical data in the current deployment interconnection area.

[0010] Preferably, whether the guiding effect of identifying environmental factors is a long-term tendency or a short-term tendency includes: Extract the target ad network area, obtain the number of user searches for the target ad within the n historical monitoring windows prior to the current time in the target ad network area, and aggregate the user search counts of the target ad according to the time granularity to construct the original search count sequence. And based on the original search count sequence Data in Formula (1) determines the relative search intensity Based on relative search strength Constructed as behavioral time-series vectors ;in, It is the number of the data in the original search count sequence, and The range of values ​​for is [1, ... ], This represents the maximum value of the data number in the original search count sequence; behavioral time vector Divided into A length of The sliding window is used, and the local coefficient of variation within each sliding window is determined by formula (2). Based on local coefficient of variation Constructing the coefficient of variation sequence And the coefficient of variation sequence is determined based on the calculation formula (3). Fluctuation entropy ;in, Let be the number of the sliding window, and The range of values ​​for is [1, ... ], It is obtained through manual settings. It is derived from the time length of the time granularity and is directly proportional to the time length of the time granularity, which can be up to 6 times the time length of the time granularity. The relative search intensity is defined based on the calculation formula (4). The first-order difference (gradient) And second-order difference (acceleration) Based on second-order difference The acceleration energy value is determined by formula (5). Used to quantify the severity of change; based on fluctuation entropy. and acceleration energy value The binary classification decision function for the target advertisement is determined by formula (6). ; The calculation formula (1) is: ; In the formula, The average number of user searches in the historical data of the target internet region, aggregated by time granularity; The standard deviation of historical user search counts aggregated by time granularity; To prevent the denominator from being divided by zero by a smoothing factor; The calculation formula (2) is: ; In the formula, For the first The standard deviation of the search intensity within a window For the first The mean of the search intensity within each window; To prevent the denominator from being divided by zero by a smoothing factor; The calculation formula (3) is: ; In the formula, For the first Local coefficient of variation In the overall coefficient of variation sequence The probability distribution density in; The calculation formula (4) is: ; The calculation formula (5) is: ; The calculation formula (6) is: ; In the formula, The stability threshold set based on historical fluctuation entropy can be a manually selected percentile value from the historical fluctuation entropy. It is used to determine whether the search behavior is maintained at a stable level. The acceleration energy threshold is set based on historical acceleration energy values, and the value can be a manually selected percentile value among the historical acceleration energy values.

[0011] Preferably, the step of delivering relevant advertisements to users entering the advertising interconnection area includes: Extract the preset delivery frequency of the target advertisement, and deliver the target advertisement in the current interconnected area based on the delivery frequency; wherein, the preset delivery frequency is set according to the investment funds of the advertising manufacturer and is proportional to the investment funds.

[0012] Preferably, the stopping time for predicting short-term tendencies based on user behavior data includes: Divide the day into several time windows according to the standard length of time. And obtain the target ads in each time window. User behavior data within; numbering the historical monitoring windows. Same time window The data is integrated into a comparison time window group, and the initial fluctuation factor is determined based on the user behavior data within the comparison time window group using formula (7). The standard time length is determined manually. For each time window of the day The number, Time window within the same period The numbering on different days, and The range of values ​​for is [1, ... ], The number of days in the historical monitoring window; Preliminary fluctuation factors for each comparison time window group corresponding to the target advertisement The characteristic fluctuation factor of the target advertisement is obtained by averaging. When the characteristic volatility factor Less than the volatility factor threshold At that time, the stopping time of the short-term tendency of the target advertisement is determined by formula (8). When the characteristic volatility factor Not less than the volatility factor threshold At the same time, continue to acquire the characteristic fluctuation factor of the target advertisement. Until the characteristic volatility factor Less than the volatility factor threshold The stopping time of the short-term tendency of the target advertisement is determined by calculation formula (8). Among them, volatility factor threshold It is based on historical characteristic fluctuation factor The value is determined and is generally taken as the historical characteristic volatility factor. The percentile value selected manually can also be 0; The calculation formula (7) is: ; In the formula, , and It is a proportional adjustment coefficient used to adjust the contribution of different behaviors to the interest judgment based on historical data, and , , and All are greater than 0; The total number of user searches for the target ad within the corresponding time window. This represents the total number of user clicks on the target ad within the corresponding time window. The total time a user spends on the target ad within the corresponding time window. hour, , , ; The calculation formula (8) is as follows: ; In the formula, For the current time, It is based on the characteristic fluctuation factor of each advertisement in historical data. Less than the volatility factor threshold The duration from the start of the day when the total number of user searches, total number of user clicks, and total dwell time for the advertisement are all less than the corresponding threshold is determined. It is generally the average of the durations mentioned above, and the corresponding thresholds are obtained through manual setting.

[0013] Preferably, dynamically delivering relevant advertisements to users entering the advertising interconnection area after the stop time includes: When the time reaches the stop time of the target advertisement's short-term tendency. When the number of times the target ad is placed in the current interconnected area is identified; when the number of times the ad is placed exceeds the number of times threshold, the daily placement frequency of the target ad is reduced to two-thirds of the previous day, until the placement frequency is less than once a day, at which point the target ad is no longer placed in the current interconnected area; wherein, the number of times threshold is the average number of times each ad is placed in the current interconnected area. When the number of times the target ad is not greater than the number of times threshold, the daily frequency of the target ad will be reduced to half of the previous day's frequency until the frequency is less than once a day, at which point the target ad will be stopped from being placed in the current interconnected area.

[0014] Preferably, the step of placing advertisements according to preset advertisements includes: The manually selected advertisements are placed in the current interconnected area according to the corresponding preset placement frequency; the preset placement frequency is set according to the investment funds of the advertising manufacturer and is proportional to the investment funds.

[0015] A second aspect of the present invention provides an environmentally aware dynamic advertising delivery system, comprising: a regional data collection module and a sensing delivery module; The regional data collection module is used to plan the interconnected advertising areas and acquire user behavior data within the interconnected advertising areas; wherein, the user behavior data includes the number of user searches for each advertisement, the number of user clicks on each advertisement, and the time users spend in the advertisement; The perception and delivery module identifies whether there are environmental factors guiding users within the interconnected delivery area based on the historical user search counts of users within the interconnected delivery area. Yes, it identifies whether the guiding effect of environmental factors is a long-term or short-term tendency; when it is a long-term tendency, it delivers relevant advertisements to users entering the advertising interconnection area; when it is a short-term tendency, it predicts the stopping time of the short-term tendency based on user behavior data, delivers relevant advertisements to users entering the advertising interconnection area before the stopping time, and dynamically delivers relevant advertisements to users entering the advertising interconnection area after the stopping time. No, the ads will be placed according to the preset ads.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention plans and acquires user behavior data within the interconnected advertising area; identifies whether environmental factors guiding users exist within the interconnected advertising area based on the historical search counts of users within the area; if yes, it identifies whether the guiding effect of the environmental factors is a long-term or short-term trend; if it is a long-term trend, it delivers relevant advertisements to users entering the interconnected advertising area; if it is a short-term trend, it predicts the stopping time of the short-term trend based on user behavior data, delivers relevant advertisements to users entering the interconnected advertising area before the stopping time, and dynamically delivers relevant advertisements to users entering the interconnected advertising area after the stopping time; otherwise, it delivers advertisements according to preset advertisements; this invention solves the technical problem of difficulty in improving the accuracy of advertisement delivery based on the perception of the surrounding environment, and improves the accuracy of advertisement delivery.

[0017] 2. The steps introduce a multi-dimensional feature recognition algorithm that combines fluctuation entropy analysis and gradient acceleration detection. This innovatively reveals the intrinsic nature of environmental factors driving user behavior from two dimensions: the stability of data changes and the intensity of mutations. It can accurately distinguish between long-term stable guidance tendencies caused by commercial atmosphere and geographical attributes and short-term explosive guidance tendencies caused by hot events and temporary promotions. This improves the sensitivity and accuracy of the advertising system to changes in complex environmental factors, thereby achieving efficient allocation and utilization of advertising resources in the dynamically changing Internet environment. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the operation steps of the present invention; Figure 2 This is a schematic diagram illustrating the operational steps of the present invention to determine whether the guiding effect is a long-term or short-term tendency; Figure 3 This is a schematic diagram of the system modules of the present invention. Detailed Implementation

[0020] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 The first aspect of this invention provides a method for dynamic advertising delivery based on environmental awareness, comprising: Plan the interconnected advertising areas and acquire user behavior data within those areas; the user behavior data includes the number of user searches for each ad, the number of user clicks on each ad, and the time users spend on the ads. Identify whether there are environmental factors guiding users within the targeted interconnection area based on the historical user search counts of users within the targeted interconnection area; Yes, it identifies whether the guiding effect of environmental factors is a long-term or short-term tendency; when it is a long-term tendency, it delivers relevant advertisements to users entering the advertising interconnection area; when it is a short-term tendency, it predicts the stopping time of the short-term tendency based on user behavior data, delivers relevant advertisements to users entering the advertising interconnection area before the stopping time, and dynamically delivers relevant advertisements to users entering the advertising interconnection area after the stopping time. No, the ads will be placed according to the preset ads.

[0022] It is worth noting that this invention provides a method and system for dynamic advertising delivery based on environmental awareness. By integrating regional data collection, user behavior data analysis, and environmental factor identification, it achieves significant improvements in the accuracy of advertising delivery. First, through dynamic planning and real-time data collection of interconnected delivery areas, this method can comprehensively understand the behavioral patterns of users within the target area, including search frequency, click count, and dwell time on different advertisements. This data not only helps advertisers adjust their delivery strategies in real time but also provides strong support for the personalization and customization of advertisements. Second, through in-depth analysis of user behavior by the perception delivery module, the system can identify potential long-term or short-term tendencies in user behavior regarding advertising content. This identification capability is significant because it can accurately infer a user's current needs or interests based on their historical behavior, thereby making advertising delivery more aligned with the user's actual needs. For example, when a user's behavior is detected to exhibit a long-term tendency, the system can deliver advertisements related to the user's long-term interests or needs based on this trend, thus improving the relevance and effectiveness of the advertisements. Conversely, when users exhibit short-term tendencies, the system can predict when other users' short-term interests will cease by analyzing the brief dwell time on ads, and push relevant ads to users within that time period. This predictive capability ensures that ads are presented to users at the most appropriate time, improving ad acceptance and conversion rates. During dynamic delivery, when the perception module fails to identify effective environmental factors or guiding effects, the system can operate according to preset ad delivery strategies. This provides advertisers with a certain degree of flexibility, ensuring that ad delivery remains effective even when personalized user behavior data is unavailable, avoiding delivery errors due to data gaps. Furthermore, by combining user behavior data with the identification of environmental factors, dynamic ad adjustment extends beyond the ad content itself to the timing and frequency of ad display. The system dynamically adjusts the ad delivery cycle based on predicted user dwell time and behavioral tendencies, avoiding excessive user disruption and improving user experience. In this process, advertising becomes not only a simple product promotion tool but also a medium for interaction with users, enhancing the affinity and effectiveness of ad delivery. In summary, this approach, through multi-dimensional data analysis and precise environmental awareness, makes the advertising delivery system more adaptable and flexible. While improving the accuracy of ad delivery, it also strengthens the interaction and relationship between ads and users, helping advertisers make optimization decisions based on real-time data and achieve more efficient marketing.

[0023] The planning for the deployment of interconnected areas in this application includes: Obtain geographic map data within the target delivery area, and divide the target delivery area into multiple geographic grids of a preset size; wherein the target delivery area is defined manually or generated based on administrative division data; Based on the geographic map data, the geographical locations and types of various infrastructures within the target deployment area are identified; the types include stadiums, shopping malls, schools, hospitals, office buildings, and communities. Based on the type of infrastructure, the corresponding influence radius parameter is invoked; taking the geographical location of the infrastructure as the center, the coverage area of ​​the corresponding infrastructure is determined according to the influence radius parameter; wherein, the influence radius parameter is determined based on the historical pedestrian density of infrastructure of the same type, for example, by dividing the average historical pedestrian density of the current type of infrastructure by the average historical pedestrian density of all infrastructures, and then multiplying the ratio by a fixed distance to obtain the influence radius parameter, the fixed distance being manually set; Traverse all geographic grids and identify whether a geographic grid falls within the coverage of one or more infrastructures. If a geographic grid falls within the coverage of only one infrastructure, mark that geographic grid as the target grid of that infrastructure. If a geographic grid falls within the coverage of multiple infrastructures, mark the infrastructure closest to the current geographic grid as the target grid of the current grid. Determine the deployment interconnection area of ​​the infrastructure based on the target grids of each infrastructure.

[0024] It's worth noting that this step involves acquiring and planning geographic map data for the target delivery area, dividing it into multiple small geographic grids. This gridded approach allows for finer-grained ad delivery, enabling more precise ad content delivery to users in each region. This method allows for delivery planning based on the influence radius of various infrastructure facilities, ensuring accurate ad coverage in areas with high user activity, avoiding wasted ad resources and ineffective delivery. Simultaneously, identifying the geographic locations and influence radii of different types of infrastructure, and calculating the influence radius using historical pedestrian density data, reflects user aggregation in different areas, thus better predicting and optimizing ad delivery strategies. For example, ad delivery accuracy is significantly improved in densely populated areas such as shopping malls, schools, or hospitals, allowing ads to attract target users more effectively. By setting the influence radius parameter, the ad coverage can be adjusted according to the actual situation of different types of infrastructure, thereby improving user experience and ad effectiveness. Furthermore, delineating interconnected delivery areas based on the identified target grids enables precise ad delivery area selection, avoiding overly large or small ad coverage areas, further improving ad targeting and conversion rates. This planning approach can significantly improve the effectiveness of advertising, enhance the accuracy of ad targeting and personalized recommendations, optimize the match between ad content and user needs, and thus effectively promote the overall performance and return on investment of the advertising system.

[0025] This application obtains user behavior data within the interconnected area, including: After user authorization, when the location device identifies the user entering the advertising interconnection area, the SDK data points and server logs are used to obtain the number of user searches for each advertisement, the number of user clicks on each advertisement, and the time the user spends in the advertisement within the advertising interconnection area.

[0026] In this application, the identification of whether environmental factors guiding users exist within the targeted interconnection area is based on the historical user search frequency of users within the targeted interconnection area, including: The system collects historical user search counts for each advertisement within the previous historical monitoring window in the target area. Advertisements are extracted sequentially and marked as target advertisements. The historical user search counts for these target advertisements are aggregated according to time granularity to construct a search popularity time series for the target advertisements. Statistical characteristics of each search popularity time series are obtained and marked as statistical characteristic one. The statistical characteristics include the mean and maximum value of user search counts, forming an initial vector describing the search behavior characteristics of the current area. The historical monitoring window is manually set, ranging from 3 to 10 days. The time granularity is manually set, and can be 1 hour or 30 minutes. The statistical characteristic is the historical user search counts included in the target advertisement within the corresponding search popularity time series. The system acquires statistical characteristics of the search popularity time sequence of the current interconnected region within the n historical monitoring windows prior to the current time, and marks these statistical characteristics as statistical characteristic two. When the deviation factor of the statistical characteristics in statistical characteristic one and statistical characteristic two is greater than the deviation threshold, it is determined that there are environmental factors in the current interconnected region that guide users to favor the target advertisement; when the deviation factor of the statistical characteristics in statistical characteristic one and statistical characteristic two is not greater than the deviation threshold, it is determined that there are no environmental factors in the current interconnected region that guide users to favor the target advertisement. Here, n is manually set and can be 3; the deviation threshold is determined based on the statistical distribution of the historically calculated deviation factors, for example, taking 90 points from the historical deviation factor sequence. The deviation factor is calculated as follows: PL = (JZ1 - JZ2) / BJZ + (DZ1 - DZ2) / BDZ; PL is the deviation factor, JZ1 is the mean of statistical feature one, JZ2 is the mean of statistical feature two, DZ1 is the maximum value of statistical feature one, DZ2 is the maximum value of statistical feature two, BJZ is the standard mean difference obtained from historical data in the current deployment interconnection area, such as the average of several historical statistical feature one means minus the corresponding statistical feature two means, and BDZ is the standard maximum value difference obtained from historical data in the current deployment interconnection area, such as the average of several historical statistical feature one maximum values ​​minus the corresponding statistical feature two maximum values.

[0027] It is worth noting that this step, through dynamic monitoring and analysis of users' historical search behavior within the target area, enables precise advertising decisions, thereby significantly improving the efficiency and effectiveness of advertising. This step collects historical user search counts for ads and aggregates them at the time granularity to construct a time series of ad search popularity, which can vividly depict the popularity of ads and user attention trends within a specific time period. By extracting statistical characteristics from the search popularity time series within the historical monitoring window, quantitative data support can be provided to determine whether environmental factors guiding user preferences exist in the current target area. When comparing statistical feature one and statistical feature two, the calculation of the deviation factor not only determines whether the current area's ad search popularity differs significantly from historical trends but also further refines the evaluation of the guiding role of environmental factors. When the deviation factor is greater than the deviation threshold, potential environmental factors within the target area can be effectively identified. These environmental factors may be a sudden increase or decrease in the popularity of a certain ad within a specific time period, thus revealing potential user preferences. Conversely, if the deviation factor is small, it indicates that user behavior within the target area remains stable. This method enables precise control over ad placement, avoiding ineffective ad delivery and ensuring that ad content and format truly align with user interests and needs, thereby maximizing click-through rates and conversion rates. Furthermore, this method is adaptable to different time periods and regions, offering strong flexibility and versatility. Based on this data-driven, environment-aware ad placement approach, it effectively improves ad targeting, enhances advertisers' ROI, and avoids ad overload and a deterioration in user experience.

[0028] It should be noted that the existence of environmental factors that guide users can also be determined through expert settings. For example, if a large concert is held in Stadium 1 during a certain period of time, it can be determined that there are environmental factors that guide users during that period of time, and that this is a short-term tendency.

[0029] It should be noted that the steps for obtaining the statistical characteristics of the search popularity time series of the advertising interconnection area within the n historical monitoring windows before the current time are as follows: collect the historical user search counts of each advertisement in the advertising interconnection area within the n historical monitoring windows before the current time, extract the advertisements in sequence, mark the advertisements as target advertisements, aggregate the historical user search counts of the target advertisements according to the time granularity to construct the search popularity time series of the target advertisements, and obtain the statistical characteristics of each search popularity time series.

[0030] Please see Figure 2 In this application, the guiding effect of environmental factors is identified as either a long-term or short-term tendency, including: Extract the target ad network area, obtain the number of user searches for the target ad within the n historical monitoring windows prior to the current time in the target ad network area, and aggregate the user search counts of the target ad according to the time granularity to construct the original search count sequence. And based on the original search count sequence Data in Formula (1) determines the relative search intensity Based on relative search strength Constructed as behavioral time-series vectors ;in, It is the number of the data in the original search count sequence, and The range of values ​​for is [1, ... ], This represents the maximum value of the data number in the original search count sequence; This step introduces fluctuation entropy to determine the stationarity of data changes. Long-term trends are usually accompanied by stable periodic or random fluctuations, with high and stable entropy values. Short-term trends (such as sudden hotspots) manifest as drastic changes in the data sequence, leading to an increase in local orderliness (orderliness refers to the randomness of the numerical distribution deviating from the normal distribution), and the entropy value will decrease significantly. If the fluctuation entropy is lower than the preset stationarity threshold, it indicates that the data sequence has an abnormal ordered structure and is judged as a suspected short-term trend; otherwise, it tends to be a long-term trend: the behavior time series vector Divided into A length of The sliding window is used, and the local coefficient of variation within each sliding window is determined by formula (2). Based on local coefficient of variation Constructing the coefficient of variation sequence And the coefficient of variation sequence is determined based on the calculation formula (3). Fluctuation entropy ;in, Let be the number of the sliding window, and The range of values ​​for is [1, ... ], It is obtained through manual settings. It is derived from the time length of the time granularity and is directly proportional to the time length of the time granularity, which can be up to 6 times the time length of the time granularity. Because a simple growth rate is insufficient to distinguish breakout points, while acceleration can effectively capture the "inflection point" of change, this step introduces gradient acceleration analysis to further confirm whether it is "always like this" or "sudden growth": The relative search intensity is defined based on the calculation formula (4). The first-order difference (gradient) And second-order difference (acceleration) Based on second-order difference The acceleration energy value is determined by formula (5). Used to quantify the severity of change; based on fluctuation entropy. and acceleration energy value The binary classification decision function for the target advertisement is determined by formula (6). ; The calculation formula (1) is: ; In the formula, The average number of user searches in the historical data of the target internet region, aggregated by time granularity; The standard deviation of historical user search counts aggregated by time granularity; To prevent the denominator from being divided by zero, the smoothing factor can be set to 1; The calculation formula (2) is: ; In the formula, For the first The standard deviation of the search intensity within a window For the first The mean of the search intensity within each window; To prevent the denominator from being divided by zero, the smoothing factor can be set to 1; The calculation formula (3) is: ; In the formula, For the first Local coefficient of variation In the overall coefficient of variation sequence The probability distribution density in; The calculation formula (4) is: ; It should be noted that when hour, , ; The calculation formula (5) is: ; The calculation formula (6) is: ; In the formula, The stability threshold set based on historical fluctuation entropy can be a manually selected percentile value from the historical fluctuation entropy. It is used to determine whether the search behavior is maintained at a stable level. The acceleration energy threshold is set based on historical acceleration energy values, and the value can be a manually selected percentile value among the historical acceleration energy values.

[0031] It is worth noting that this step, by introducing a multi-dimensional feature recognition algorithm combining fluctuation entropy analysis and gradient acceleration detection, enhances the sensitivity and accuracy of the advertising delivery system to changes in complex environmental factors, thereby achieving efficient allocation and utilization of advertising resources in the dynamically changing internet environment. This method abandons the traditional, crude judgment logic that relies solely on a single search volume growth threshold. It innovatively starts from two dimensions: the stability of data changes and the intensity of sudden changes, profoundly revealing the intrinsic nature of environmental factors driving user behavior. It can accurately distinguish between long-term stable guidance tendencies constituted by commercial atmosphere and geographical attributes, and short-term explosive guidance tendencies caused by hot events and temporary promotions. By calculating temporal fluctuation entropy, the system can effectively capture changes in the orderliness of data sequences, eliminate random noise interference, and accurately identify abnormal ordered structures brought about by short-term hotspots, thus avoiding the error of misjudging instantaneous traffic as long-term demand. Simultaneously, energy analysis based on gradient acceleration can keenly capture the inflection point and explosive intensity of search behavior growth, ensuring a rapid response in the early stages of traffic surges. This dual verification mechanism greatly reduces the false positive and false negative rates, enabling the system to formulate differentiated delivery strategies for environmental factors of different natures. For long-term trends, the system can reliably execute continuous advertising strategies, deeply explore the potential commercial value within a region, maintain consistent brand exposure, and cultivate long-term consumer habits. For short-term trends, the system can promptly switch to a time-sensitive advertising mode, maximizing ad reach during periods of high traffic by accurately predicting when to stop, and quickly adjusting or stopping advertising when traffic declines, effectively avoiding ineffective waste of advertising resources and budget overruns. Furthermore, this refined trend identification not only improves ad conversion rates and ROI but also optimizes user experience by reducing unnecessary disruptions, enhancing user acceptance of ad content. Ultimately, it achieves the dual goals of maximizing advertiser revenue and improving user satisfaction, providing solid technical support for the intelligent and precise development of the entire advertising ecosystem.

[0032] It should be noted that if the acceleration energy value If the acceleration energy is not greater than the acceleration energy threshold, it is considered low acceleration energy, indicating that the search behavior remains at a stable level without drastic changes and tends to be long-term. If the acceleration energy value If the acceleration energy is greater than the acceleration energy threshold, it is considered high acceleration energy, indicating that the search behavior has recently experienced a sharp increase in acceleration and tends to be short-term.

[0033] This application includes delivering relevant advertisements to users entering the aforementioned interconnected area, including: Extract the preset delivery frequency of the target advertisement, and deliver the target advertisement in the current interconnected area based on the delivery frequency; wherein, the preset delivery frequency is set according to the investment funds of the advertising manufacturer and is proportional to the investment funds.

[0034] In this application, the stopping time for predicting short-term tendencies based on user behavior data includes: Divide the day into several time windows according to the standard length of time. And obtain the target ads in each time window. User behavior data within; numbering the historical monitoring windows. Same time window The data is integrated into a comparison time window group, and the initial fluctuation factor is determined based on the user behavior data within the comparison time window group using formula (7). The standard time length is set manually and can be 30 minutes. For each time window of the day The numbering, such as the numbering of 0:00-0:30 on a certain day when the standard time length is 30 minutes. Numbering from 0:30 to 1:00 ; Time window within the same period The numbering on different days, and The range of values ​​for is [1, ... ], This refers to the number of days in the historical monitoring window. The historical monitoring window is manually set and ranges from 3 to 10 days. For example, the first day within the historical monitoring window is the time period from 14:00 to 14:30. The number of the second day ; Preliminary fluctuation factors for each comparison time window group corresponding to the target advertisement The characteristic fluctuation factor of the target advertisement is obtained by averaging. When the characteristic volatility factor Less than the volatility factor threshold At that time, the stopping time of the short-term tendency of the target advertisement is determined by formula (8). When the characteristic volatility factor Not less than the volatility factor threshold At the same time, continue to acquire the characteristic fluctuation factor of the target advertisement. Until the characteristic volatility factor Less than the volatility factor threshold The stopping time of the short-term tendency of the target advertisement is determined by calculation formula (8). Among them, volatility factor threshold It is based on historical characteristic fluctuation factor The value is determined and is generally taken as the historical characteristic volatility factor. The percentile value selected manually can also be 0; The calculation formula (7) is: ; In the formula, , and It is a proportional adjustment coefficient used to adjust the contribution of different behaviors to the interest judgment based on historical data, and , , and All are greater than 0; The total number of user searches for the target ad within the corresponding time window. This represents the total number of user clicks on the target ad within the corresponding time window. The total time a user spends on the target ad within the corresponding time window. hour, , , ; The calculation formula (8) is as follows: ; In the formula, For the current time, It is based on the characteristic fluctuation factor of each advertisement in historical data. Less than the volatility factor threshold The duration from the start of the day when the total number of user searches, total number of user clicks, and total dwell time for the advertisement are all less than the corresponding threshold is determined. It is generally the average of the durations mentioned above, and the corresponding thresholds are obtained through manual setting.

[0035] It's worth noting that this step, by predicting the cessation time of short-term trends using user behavior data, significantly improves the dynamic responsiveness of ad delivery. By precisely analyzing user behavior data such as ad searches, clicks, and dwell times at different times of the day, the system can effectively identify trends in ad popularity, especially sudden short-term trends. This time-window-based analysis method can capture the regularity of ad popularity fluctuations by comparing data from multiple historical monitoring windows within the same time period. By calculating preliminary fluctuation factors and characteristic fluctuation factors, the system can identify whether ads exhibit continuous and stable user interest or sudden interest that only appears in the short term. If the characteristic fluctuation factor is lower than the set fluctuation factor threshold, the system will accurately predict the cessation time of short-term ad trends based on these trends. This mechanism not only helps the system to promptly identify trending ads and enhance ad exposure in a short period, but also avoids the continuous delivery of ineffective ads, ensuring the efficient use of the advertising budget. In practice, the calculation of the characteristic fluctuation factor can be dynamically adjusted based on the specific performance of the advertisement, thereby avoiding poor results due to excessive or insufficient advertising. Through this mechanism, the advertising system can not only respond flexibly to sudden trending topics but also ensure the continuity of advertising during stable periods, thus improving the accuracy and real-time nature of advertising. This method provides advertisers with highly targeted advertising strategies while optimizing the user experience, avoiding duplicate and ineffective ad pushes, and increasing user acceptance and engagement. Based on this approach, advertisers can accurately capture short-term trending topics, improving conversion rates and ROI, while the platform can enhance the stability and long-term benefits of advertising through real-time data monitoring and dynamic adjustments.

[0036] In this application, relevant advertisements are dynamically delivered to users entering the advertising interconnection area after a stop time, including: When the time reaches the stop time of the target advertisement's short-term tendency. When the number of times the target ad is placed in the current interconnected area is identified; when the number of times the ad is placed exceeds the number of times threshold, the daily placement frequency of the target ad is reduced to two-thirds of the previous day, until the placement frequency is less than once a day, at which point the target ad is no longer placed in the current interconnected area; wherein, the number of times threshold is the average number of times each ad is placed in the current interconnected area. When the number of times the target ad is not greater than the number of times threshold, the daily frequency of the target ad will be reduced to half of the previous day's frequency until the frequency is less than once a day, at which point the target ad will be stopped from being placed in the current interconnected area.

[0037] In this application, advertising is delivered according to a preset advertising plan, including: The manually selected advertisements are placed in the current interconnected area according to the corresponding preset placement frequency; the preset placement frequency is set according to the investment funds of the advertising manufacturer and is proportional to the investment funds.

[0038] Please see Figure 3 A second aspect of the present invention provides an environmentally-aware dynamic advertising delivery system, comprising: a regional data collection module and a perception delivery module; The regional data collection module is used to plan the interconnected advertising areas and acquire user behavior data within the interconnected advertising areas; wherein, the user behavior data includes the number of user searches for each advertisement, the number of user clicks on each advertisement, and the time users spend in the advertisement; The perception and delivery module identifies whether there are environmental factors guiding users within the interconnected delivery area based on the historical user search counts of users within the interconnected delivery area. Yes, it identifies whether the guiding effect of environmental factors is a long-term or short-term tendency; when it is a long-term tendency, it delivers relevant advertisements to users entering the advertising interconnection area; when it is a short-term tendency, it predicts the stopping time of the short-term tendency based on user behavior data, delivers relevant advertisements to users entering the advertising interconnection area before the stopping time, and dynamically delivers relevant advertisements to users entering the advertising interconnection area after the stopping time. No, the ads will be placed according to the preset ads.

[0039] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0040] Working principle of the invention: This invention plans the targeting of interconnected regions and acquires user behavior data within these regions. It identifies whether environmental factors guiding users exist within the interconnected regions based on the historical search counts of users within those regions. When such environmental factors exist, it identifies whether the guiding effect of the corresponding advertisements is a long-term or short-term trend. If it is a long-term trend, relevant advertisements are displayed to users entering the interconnected regions. If it is a short-term trend, the invention predicts the stopping time of the short-term trend based on user behavior data, displays relevant advertisements to users entering the interconnected regions before the stopping time, and dynamically displays relevant advertisements to users entering the interconnected regions after the stopping time. When no environmental factors guiding users exist within the interconnected regions, advertisements are displayed according to preset advertisements.

[0041] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for dynamic advertising delivery based on environmental perception, characterized in that, include: Plan the interconnected advertising areas and acquire user behavior data within those areas; the user behavior data includes the number of user searches for each ad, the number of user clicks on each ad, and the time users spend on the ads. Identify whether there are environmental factors guiding users within the targeted interconnection area based on the historical user search counts of users within the targeted interconnection area; Yes, it identifies whether the guiding effect of environmental factors is a long-term or short-term tendency; when it is a long-term tendency, it delivers relevant advertisements to users entering the advertising interconnection area; when it is a short-term tendency, it predicts the stopping time of the short-term tendency based on user behavior data, delivers relevant advertisements to users entering the advertising interconnection area before the stopping time, and dynamically delivers relevant advertisements to users entering the advertising interconnection area after the stopping time. No, the ads will be placed according to the preset ads.

2. The method for dynamic advertising delivery based on environmental perception according to claim 1, characterized in that, The planning for the deployment of interconnected areas includes: Obtain geographic map data within the target delivery area, and divide the target delivery area into multiple geographic grids of a preset size; wherein the target delivery area is defined manually or generated based on administrative division data; Based on the geographic map data, the geographical locations and types of various infrastructures within the target deployment area are identified; the types include stadiums, shopping malls, schools, hospitals, office buildings, and communities; Based on the type of infrastructure, the corresponding influence radius parameter is invoked; taking the geographical location of the infrastructure as the center, the coverage area of ​​the corresponding infrastructure is determined according to the influence radius parameter; wherein, the influence radius parameter is determined based on the historical pedestrian density of infrastructure of the same type; Traverse all geographic grids and identify whether a geographic grid falls within the coverage of one or more infrastructures. If a geographic grid falls within the coverage of only one infrastructure, mark that geographic grid as the target grid of that infrastructure. If a geographic grid falls within the coverage of multiple infrastructures, mark the infrastructure closest to the current geographic grid as the target grid of the current grid. Determine the deployment interconnection area of ​​the infrastructure based on the target grids of each infrastructure.

3. The method for dynamic advertising delivery based on environmental perception according to claim 1, characterized in that, The acquisition of user behavior data within the interconnected area includes: After user authorization, when the location device identifies the user entering the advertising interconnection area, the SDK data points and server logs are used to obtain the number of user searches for each advertisement, the number of user clicks on each advertisement, and the time the user spends in the advertisement within the advertising interconnection area.

4. The method for dynamic advertising delivery based on environmental perception according to claim 1, characterized in that, The step of identifying whether there are environmental factors guiding users within the targeted interconnection area based on the historical user search frequency of users within the targeted interconnection area includes: The system collects the historical user search counts of each advertisement in the interconnected region within the previous historical monitoring window at the current time, extracts the advertisements sequentially, and marks them as target advertisements. It then aggregates the historical user search counts of the target advertisements according to time granularity to construct a search popularity time series for the target advertisements. Statistical features of each search popularity time series are obtained and marked as statistical feature one. The statistical features include the mean and maximum value of user search counts. Specifically, the statistical feature is the historical user search count included in the target advertisement within the corresponding search popularity time series. The system acquires statistical characteristics of the search popularity time sequence of the current interconnected region within the n historical monitoring windows prior to the current time, and marks these statistical characteristics as statistical characteristic two. When the deviation factor of the statistical characteristics in statistical characteristic one and statistical characteristic two is greater than the deviation threshold, it is determined that there are environmental factors in the current interconnected region that guide users to favor the target advertisement. When the deviation factor of the statistical characteristics in statistical characteristic one and statistical characteristic two is not greater than the deviation threshold, it is determined that there are no environmental factors in the current interconnected region that guide users to favor the target advertisement. The deviation threshold is determined based on the statistical distribution of the deviation factors calculated historically.

5. The method for dynamic advertising delivery based on environmental perception according to claim 4, characterized in that, Whether the guiding role of environmental factors is identified as a long-term or short-term tendency includes: Extract the target ad network area, obtain the number of user searches for the target ad within the n historical monitoring windows prior to the current time in the target ad network area, and aggregate the user search counts of the target ad according to the time granularity to construct the original search count sequence. And based on the original search count sequence Data in Formula (1) determines the relative search intensity Based on relative search strength Constructed as behavioral time-series vectors ;in, It is the number of the data in the original search count sequence, and The range of values ​​for is [1, ... ], This represents the maximum value of the data number in the original search count sequence; behavioral time vector Divided into A length of The sliding window is used, and the local coefficient of variation within each sliding window is determined by formula (2). Based on local coefficient of variation Constructing the coefficient of variation sequence And the coefficient of variation sequence is determined based on the calculation formula (3). Fluctuation entropy ;in, Let be the number of the sliding window, and The range of values ​​for is [1, ... ], It is obtained based on the time length of the time granularity; The relative search intensity is defined based on the calculation formula (4). First-order difference and second-order difference Based on second-order difference The acceleration energy value is determined by formula (5). Based on fluctuation entropy and acceleration energy value The binary classification decision function for the target advertisement is determined by formula (6). ; The calculation formula (1) is: ； In the formula, The average number of user searches in the historical data of the target internet region, aggregated by time granularity; The standard deviation of historical user search counts aggregated by time granularity; To prevent the denominator from being divided by zero by a smoothing factor; The calculation formula (2) is: ； In the formula, For the first The standard deviation of the search intensity within a window For the first The mean of the search intensity within each window; To prevent the denominator from being divided by zero by a smoothing factor; The calculation formula (3) is: ； In the formula, For the first Local coefficient of variation In the overall coefficient of variation sequence The probability distribution density in; The calculation formula (4) is: ； The calculation formula (5) is: ； The calculation formula (6) is: ； In the formula, A stability threshold set based on historical fluctuation entropy; It is used to determine whether the search behavior is maintained at a stable level. The acceleration energy threshold is set based on historical acceleration energy values.

6. The method for dynamic advertising delivery based on environmental perception according to claim 4, characterized in that, The delivery of relevant advertisements to users entering the advertising interconnection area includes: Extract the preset delivery frequency of the target advertisement, and deliver the target advertisement in the current interconnected area based on the delivery frequency; wherein, the preset delivery frequency is set according to the investment funds of the advertising manufacturer and is proportional to the investment funds.

7. The method for dynamic advertising delivery based on environmental perception according to claim 4, characterized in that, The stopping time based on user behavior data to predict short-term tendencies includes: Divide the day into several time windows according to the standard length of time. And obtain the target ads in each time window. User behavior data within; numbering the historical monitoring windows. Same time window The data is integrated into a comparison time window group, and the initial fluctuation factor is determined based on the user behavior data within the comparison time window group using formula (7). The standard time length is determined manually. For each time window of the day The number, Time window within the same period The numbering on different days, and The range of values ​​for is [1, ... ], The number of days in the historical monitoring window; Preliminary fluctuation factors for each comparison time window group corresponding to the target advertisement The characteristic fluctuation factor of the target advertisement is obtained by averaging. When the characteristic volatility factor Less than the volatility factor threshold At that time, the stopping time of the short-term tendency of the target advertisement is determined by formula (8). When the characteristic volatility factor Not less than the volatility factor threshold At the same time, continue to acquire the characteristic fluctuation factor of the target advertisement. Until the characteristic volatility factor Less than the volatility factor threshold The stopping time of the short-term tendency of the target advertisement is determined by calculation formula (8). Among them, volatility factor threshold It is based on historical characteristic fluctuation factor It is confirmed; The calculation formula (7) is: ； In the formula, , and It is a proportional adjustment coefficient used to adjust the contribution of different behaviors to the interest judgment based on historical data, and , , and All are greater than 0; The total number of user searches for the target ad within the corresponding time window. This represents the total number of user clicks on the target ad within the corresponding time window. The total time a user spends on the target ad within the corresponding time window. hour, , , ; The calculation formula (8) is as follows: ； In the formula, For the current time, It is based on the characteristic fluctuation factor of each advertisement in historical data. Less than the volatility factor threshold The duration from the start of the day until the total number of user searches, total number of user clicks, and total dwell time for the advertisement are all less than the corresponding thresholds is determined.

8. The method for dynamic advertising delivery based on environmental perception according to claim 4, characterized in that, The step of dynamically delivering relevant advertisements to users entering the interconnected advertising area after the stop time includes: When the time reaches the stop time of the target advertisement's short-term tendency. When the number of times the target ad is placed in the current interconnected area is identified; when the number of times the ad is placed exceeds the number of times threshold, the daily placement frequency of the target ad is reduced to two-thirds of the previous day, until the placement frequency is less than once a day, at which point the target ad is no longer placed in the current interconnected area; wherein, the number of times threshold is the average number of times each ad is placed in the current interconnected area. When the number of times the target ad is not greater than the number of times threshold, the daily frequency of the target ad will be reduced to half of the previous day's frequency until the frequency is less than once a day, at which point the target ad will be stopped from being placed in the current interconnected area.

9. The method for dynamic advertising delivery based on environmental perception according to claim 1, characterized in that, The method of placing advertisements according to preset advertisements includes: The manually selected advertisements are placed in the current interconnected area according to the corresponding preset placement frequency; the preset placement frequency is set according to the investment funds of the advertising manufacturer and is proportional to the investment funds.

10. An environment-aware dynamic advertising delivery system, used to run the environment-aware dynamic advertising delivery method according to any one of claims 1 to 9, characterized in that, include: Regional data collection module and sensing and delivery module; The regional data collection module is used to plan the interconnected advertising areas and acquire user behavior data within the interconnected advertising areas; wherein, the user behavior data includes the number of user searches for each advertisement, the number of user clicks on each advertisement, and the time users spend in the advertisement; The perception and delivery module identifies whether there are environmental factors guiding users within the interconnected delivery area based on the historical user search counts of users within the interconnected delivery area. Yes, it identifies whether the guiding effect of environmental factors is a long-term or short-term tendency; when it is a long-term tendency, it delivers relevant advertisements to users entering the advertising interconnection area; when it is a short-term tendency, it predicts the stopping time of the short-term tendency based on user behavior data, delivers relevant advertisements to users entering the advertising interconnection area before the stopping time, and dynamically delivers relevant advertisements to users entering the advertising interconnection area after the stopping time. No, the ads will be placed according to the preset ads.

Images