Cross-border advertisement accurate putting method, system and device based on user portraits

Abstract

The invention discloses a cross-border advertisement accurate delivery method, system and device based on a user portrait, relates to the technical field of cross-border digital marketing, and discloses the cross-border advertisement accurate delivery method, system and device based on the user portrait, which generates an accurate user portrait by integrating cross-regional behavior data. An advertisement putting strategy is dynamically calculated in combination with exchange rate, tax and logistics economic parameters, and compliance adaptation is performed based on legal and cultural rules, so that the problems of single data dimension, inaccurate user portraits and lack of economic variables in the prior art are solved; the method has the advantages that the cross-border advertisement putting accuracy and conversion effect are improved, and the compliance risk is reduced.

Description

Technical Field

[0001] The present application relates to the field of cross-border digital marketing technology, and in particular to a method, system, and device for precise cross-border advertising delivery based on user portraits. Background Art

[0002] Cross-border advertising refers to targeted marketing campaigns targeting user groups in different countries or regions. Current cross-border advertising technology suffers from three key limitations: First, in terms of data collection, existing solutions typically rely solely on users' browsing history, search history, or basic location information, failing to effectively integrate comprehensive data such as users' cross-regional behavior, multi-currency transaction characteristics, and device network environments. Second, in terms of user profile construction, traditional methods lack the ability to identify cross-border-specific characteristics such as activity patterns across time zones and multilingual interaction preferences. This results in the generated user profiles failing to accurately reflect users' true needs in cross-border consumption scenarios. Third, in the calculation of economic parameters, existing technologies generally ignore the impact of cross-border economic variables such as exchange rate fluctuations, tariff policy differences, and logistics cost fluctuations on advertising effectiveness. Furthermore, existing systems often use a single template when generating ad content, failing to dynamically adapt to the laws, regulations, religious, and cultural taboos of the target market, which can easily lead to compliance risks and cultural conflicts. These technical shortcomings collectively contribute to the current problems of low cross-border advertising accuracy, poor conversion rates, and high compliance risks. Addressing these issues, existing technologies urgently need to be improved.

[0003] The above content is only used to assist in understanding the technical solution of this application and does not constitute an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of this application is to provide a method, system and device for precise cross-border advertising delivery based on user portraits, aiming to improve the accuracy and conversion effect of cross-border advertising delivery and reduce compliance risks.

[0005] To achieve the above objectives, this application proposes a method for precise cross-border advertising delivery based on user profiles, the method comprising:

[0006] Obtain user behavior data, transaction data, and device feature data;

[0007] Perform cross-regional feature fusion processing on the behavioral data, transaction data, and device feature data to generate cross-border user profiles;

[0008] Determine the matching degree between the advertisement and the user based on the cross-border user profile, and obtain exchange rate fluctuation data, tariff rate data, and logistics cost data for the target region;

[0009] Calculate advertising economic parameters based on the matching degree between advertisements and users, exchange rate fluctuation data, tariff rate data, and logistics cost data;

[0010] Generate original ads based on cross-border user profiles and advertising economic parameters. Filter and convert the original ads according to the legal and cultural databases of the target region to generate the final ads.

[0011] The final advertisement is matched to an advertisement server cluster in a target area for advertisement delivery.

[0012] In one embodiment, the step of obtaining user behavior data, transaction data, and device feature data includes:

[0013] Collecting the behavioral data through the browser API; the behavioral data includes the user's page dwell time, click behavior sequence and language preference in multiple countries or regions;

[0014] Obtain the transaction data through the payment gateway interface; the transaction data includes transaction currency, transaction amount and commodity category;

[0015] The device characteristic data is obtained through the terminal device used by the user; the device characteristic data includes the device time zone, IP address hopping frequency and network proxy characteristics.

[0016] In one embodiment, the step of performing cross-regional feature fusion processing on the behavior data, transaction data, and device feature data to generate a cross-border user profile includes:

[0017] De-identifying the user's page dwell time and language preferences in multiple countries or regions to generate a user behavior link code containing multiple language preferences;

[0018] Correlating and matching the transaction data with the click behavior sequence to generate a user product category preference label;

[0019] Extract user activity characteristics across time zones based on device time zones, IP address hopping frequency, and network proxy features;

[0020] The user behavior link code, user product category preference label, and user cross-time zone active features are input into a preset federated learning model to generate a cross-border user profile.

[0021] In one embodiment, the step of obtaining exchange rate data, tariff data, and logistics cost data of the target region, and calculating advertising economic parameters based on the exchange rate data, tariff data, and logistics cost data includes:

[0022] Obtain real-time exchange rate fluctuation data through the central bank interface of the target region;

[0023] Call the target country's General Administration of Customs API to obtain the tariff rate data corresponding to the HS code;

[0024] Obtain logistics cost data through the logistics service provider interface; the logistics cost data includes real-time freight and estimated delivery time.

[0025] In one embodiment, the step of calculating the economic parameters of advertising delivery based on the matching degree between advertisements and users, exchange rate fluctuation data, tariff rate data, and logistics cost data specifically includes:

[0026] The economic parameters of advertising are calculated according to the following formula:

[0027] ;

[0028] in, Indicates the economic parameters of advertising placement; Indicates the matching degree between advertisement and user; Indicates exchange rate fluctuation data; Represents tariff rate data; Indicates real-time shipping rates; Indicates the estimated delivery time; represents the exchange rate gain weight; represents the tariff loss weight; Indicates the logistics freight weight; Represents the logistics timeliness weight.

[0029] In one embodiment, the steps of generating an original advertisement based on cross-border user profiles and advertising economic parameters, filtering and formatting the original advertisement according to the legal and cultural libraries of the target region, and generating a final advertisement include:

[0030] Based on the economic parameters of advertising and the product category preference tags in the cross-border user profile, the corresponding advertising template is called from the preset advertising material library, and the original advertisement is generated in combination with the economic parameters of advertising.

[0031] Parse the advertising review rules in the target region's legal database and perform sensitive word detection and prohibited content filtering on original ads;

[0032] The compliance of the visual elements of the advertisement is checked against the taboo color library, religious symbol library and festival customs library in the cultural library of the target area to generate the final advertisement.

[0033] In one embodiment, the step of matching the final advertisement to an advertisement server cluster in a target area for advertisement delivery includes:

[0034] Adjust the final ad format parameters based on the target region's device resolution distribution data, including video bitrate, multilingual subtitle embedding location, and interactive button size;

[0035] The final advertisement is converted into a standardized format that meets the requirements of the ad server cluster in the target region, and distributed to the server nodes in the target time period that meets the user's active characteristics across time zones.

[0036] In one embodiment, after the step of matching the final advertisement to an advertisement server cluster in a target area for advertisement delivery, the method further comprises:

[0037] Collect the final ad impressions, click-through rates, and conversion rates;

[0038] Adjust the weight of the user's cross-time zone active feature of the federated learning model based on the exposure, click-through rate, and conversion rate to generate an advertising optimization coefficient;

[0039] The exposure volume is matched with the number of active users in the target area during the time period using a decay function, the click-through rate and the product category preference label are subjected to cosine similarity analysis, and the conversion rate and the economic parameters of the advertising are subjected to Pearson correlation verification to generate a multi-dimensional evaluation matrix;

[0040] The final advertisement is optimized according to the advertisement optimization coefficient and the multi-dimensional evaluation matrix.

[0041] In addition, to achieve the above objectives, this application also proposes a cross-border advertising precision delivery system based on user portraits, which includes:

[0042] A first data acquisition module is used to acquire user behavior data, transaction data and device feature data;

[0043] A cross-border user profile generation module is used to perform cross-regional feature fusion processing on the behavioral data, transaction data, and device feature data to generate a cross-border user profile;

[0044] A second data acquisition module is used to determine the matching degree between the advertisement and the user based on the cross-border user profile, and to obtain exchange rate fluctuation data, tariff rate data, and logistics cost data of the target region;

[0045] Advertisement delivery economic parameter calculation module, used to calculate the economic parameters of advertisement delivery based on the matching degree between advertisement and user, exchange rate fluctuation data, tariff rate data and logistics cost data;

[0046] The final ad generation module is used to generate original ads based on cross-border user profiles and economic parameters of advertising placement, and to filter and format the original ads according to the legal and cultural databases of the target region to generate the final ads;

[0047] The advertisement delivery module is used to match the final advertisement to the advertisement server cluster in the target area for advertisement delivery.

[0048] In addition, to achieve the above-mentioned purpose, the present application also proposes a device for cross-border advertising precision delivery based on user portraits, which includes: a memory, a processor, and a cross-border advertising precision delivery program based on user portraits stored on the memory and runnable on the processor, and the cross-border advertising precision delivery program based on user portraits is configured to implement the steps of the cross-border advertising precision delivery method based on user portraits.

[0049] The method, system and device for precise cross-border advertising delivery based on user portraits proposed in this application generate precise user portraits by integrating cross-regional behavioral data, dynamically calculate advertising delivery strategies based on exchange rates, tariffs and logistics economic parameters, and perform compliance adaptation based on legal and cultural rules, thereby solving the problems of single data dimension, inaccurate user portraits and missing economic variables in the existing technology, and has the advantages of improving the accuracy of cross-border advertising delivery, conversion effects and reducing compliance risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0052] Figure 1 A flowchart illustrating an embodiment of a method for precise cross-border advertising delivery based on user portraits in this application;

[0053] Figure 2 For this application Figure 1 Detailed flow chart of step S100;

[0054] Figure 3 For this application Figure 1 Detailed flow diagram of step S200;

[0055] Figure 4 For this application Figure 1 Detailed flowchart of step S300;

[0056] Figure 5 For this application Figure 1 Detailed flowchart of step S500;

[0057] Figure 6 For this application Figure 1 Detailed flowchart of step S600;

[0058] Figure 7 A flowchart illustrating another embodiment of the cross-border advertising precision delivery method based on user portraits provided by this application;

[0059] Figure 8 A structural diagram of an embodiment of a cross-border advertising precision delivery system based on user portraits provided by this application;

[0060] Figure 9 This is a structural diagram of an embodiment of a device for precise cross-border advertising delivery based on user portraits provided in this application.

[0061] Description of Figure Numbers:

[0062] 10. Cross-border advertising precision delivery system based on user portrait; 100. First data acquisition module; 200. Cross-border user portrait generation module; 300. Second data acquisition module; 400. Advertising delivery economic parameter calculation module; 500. Final advertisement generation module; 600. Advertising delivery module; 20. Memory; 30. Processor.

[0063] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0064] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only some embodiments of this application, rather than all embodiments. The components of this application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents selected embodiments of the application. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0065] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0066] In existing technologies, cross-border advertising relies primarily on users' geographic location or single-dimensional browsing history for targeting. Due to significant differences in user behavior patterns and consumption preferences across different regions, advertising based on simple location information struggles to accurately capture users' true needs. For example, when targeting users in multiple regions, multinational e-commerce platforms may fail to consider the impact of exchange rate fluctuations on product pricing, resulting in a mismatch between advertising content and users' actual willingness to pay. Furthermore, the failure to integrate device feature data may make it impossible to identify users' true locations, resulting in pushed ad formats being incompatible with the resolution of local terminal devices.

[0067] In order to solve the above problems, those skilled in the art realized the need to build a multi-dimensional user portrait and introduce an economic factor optimization model. Through analysis, it was found that the existing technology did not associate user behavior data with device characteristics across regions, and was unable to identify the active characteristics of users when switching between multiple regions; at the same time, the advertising delivery strategy did not dynamically combine economic parameters such as exchange rates and tariffs, resulting in an imbalance between costs and benefits. Based on this, the inventor proposed to integrate user behavior link coding with device characteristics and establish an economic parameter calculation model, so as to simultaneously achieve content matching improvement and delivery cost optimization in the advertising generation stage.

[0068] Therefore, this application proposes a method for accurate cross-border advertising based on user portraits. Figure 1 The method includes steps S100 to S600, wherein:

[0069] Step S100, obtaining user behavior data, transaction data and device feature data;

[0070] Step S200: performing cross-regional feature fusion processing on the behavior data, transaction data, and device feature data to generate a cross-border user profile;

[0071] Step S300: determining the matching degree between the advertisement and the user based on the cross-border user profile, and obtaining exchange rate fluctuation data, tariff rate data, and logistics cost data of the target region;

[0072] Step S400, calculating advertising economic parameters based on the matching degree between the advertisement and the user, exchange rate fluctuation data, tariff rate data, and logistics cost data;

[0073] Step S500: Generate an original advertisement based on the cross-border user profile and advertising economic parameters, and filter and convert the original advertisement according to the legal and cultural databases of the target region to generate a final advertisement.

[0074] Step S600: Match the final advertisement to an advertisement server cluster in a target area for advertisement delivery.

[0075] In this embodiment, cross-regional feature fusion processing refers to de-identifying behavioral data from multiple countries or regions to generate a user behavior link code that includes multi-language preferences. For example, a federated learning model can be used to encrypt and calculate user behavior data distributed across different servers, retaining the user's active characteristics across time zones while avoiding data privacy leaks. Advertising economic parameters refer to quantitative indicators formed by comprehensively combining advertising matching with economic factors in the target area. Specifically, the exchange rate fluctuation data, tariff rate data, and logistics cost data can be weighted by pre-set weight coefficients to dynamically evaluate the benefit-risk ratio of advertising. Legal libraries and cultural libraries refer to databases that store advertising compliance rules and visual taboos in the target area, such as parsing sensitive word rules in legal texts through natural language processing technology, or detecting religious symbol taboos in advertising materials through image recognition algorithms.

[0076] Specifically, user behavioral data such as page dwell time and click sequences generated across multiple regions is de-identified to form a behavioral chain code that reflects the user's true interests and preferences. Device time zone and IP hopping frequency data are used to extract user activity characteristics across time zones, such as identifying patterns of users frequently using European IP addresses at night in the GMT+8 time zone. Product categories in transaction data are correlated with click behavior to generate product category preference tags, forming the core dimension of user profiles. When calculating economic parameters for advertising, exchange rate fluctuations can influence the accuracy of product price conversions in real time. Tariff rate data determines whether customs clearance costs are within the user's acceptable range. Logistics costs are linked to the impact of estimated delivery times on user purchasing decisions. During the final ad generation stage, a prohibited goods keyword filtering mechanism in the legal database removes product descriptions that do not comply with local import and export controls. Festival and customs data in the cultural database adjusts ad visual elements to align with the cultural customs of the target region.

[0077] Compared to existing technologies, existing solutions typically process user behavior data and economic environment parameters independently, resulting in a mismatch between advertising content and users' purchasing power. For example, traditional methods might recommend high-tariff products based solely on user click history, without incorporating tariff data into the actual price calculation, resulting in high click-through rates but low conversion rates. This solution integrates multi-regional features through a federated learning model, identifying potential consumer demand across multiple regions. It also incorporates economic parameters into its matching calculations, ensuring that recommended ads meet user expectations in terms of price sensitivity and logistics timeliness.

[0078] Through the above technical solution, this application can accurately identify the real needs of users in multi-region switching and dynamically adjust the price information and delivery plan in the advertising content. For example, when the user's device display frequently switches between Chinese, Japanese and Korean IP addresses, the system automatically matches the product version with the lowest tariff rate in the three countries and highlights the cross-border free shipping information in the advertisement. At the same time, the legal library filtering mechanism avoids pushing advertisements for products that are prohibited from import in the target area, reducing the risk of violations. The final generated advertising format adapts to the resolution of mainstream devices in the target area, ensuring the display effect of advertising materials and improving the user interaction experience.

[0079] In one possible implementation, reference Figure 2 , step S100 includes steps S110 to S130, wherein:

[0080] Step S110: collecting the behavior data through the browser API; the behavior data includes the user's page dwell time, click behavior sequence, and language preference in multiple countries or regions;

[0081] Step S120: Acquire the transaction data through the payment gateway interface; the transaction data includes the transaction currency, transaction amount, and commodity category;

[0082] Step S130, obtaining the device characteristic data through the terminal device used by the user; the device characteristic data includes the device time zone, IP address hopping frequency and network proxy characteristics.

[0083] In this embodiment, the browser API collects behavioral data, which refers to the use of the application program interface provided by the browser to capture the user's interactive information on the web page. Specifically, it can be implemented using a JavaScript event monitoring mechanism, such as tracking the page dwell time and click event sequence through the addEventListener function. The payment gateway interface obtains transaction data, which refers to establishing a data channel with a third-party payment platform to extract transaction records. Specifically, the OAuth 2.0 protocol can be used for authorized access to obtain encrypted transaction currency, amount, and product category information. The terminal device obtains device feature data, which refers to parsing device properties through the underlying interface of the operating system or the network protocol. Specifically, the device time zone can be read through the Settings.System class provided by the Android system, and the IP address hopping frequency and proxy server identification can be analyzed through TCP / IP data packets.

[0084] Specifically, behavioral data is collected in real time through the browser API, which can cover users' dynamic interactive behaviors on pages in multiple regions. For example, the difference in the length of time users stay on pages in Europe and the click sequence of pages in Asia can reflect cross-regional preferences. Transaction data is obtained through the payment gateway interface to ensure the integrity and timeliness of transaction currency and product category information. For example, records of users purchasing electronic products with US dollars and clothing products with euros can be associated with consumption habits in different regions. Device feature data is directly extracted through the terminal device, and the device time zone can be used to determine the user's permanent residence. The IP address jump frequency and network proxy characteristics can be used to identify whether the user uses cross-border network access tools. For example, the frequency of IP address changes caused by VPN jumps is higher than that of regular users.

[0085] Compared with existing technologies, existing cross-border advertising data collection typically relies on single geographic location information or static browsing history. For example, determining a user's country solely by IP address fails to capture multi-page interaction behaviors and differences in transaction preferences. This solution, however, integrates multi-dimensional dynamic data sources, such as associating page dwell time with transaction currency, to more accurately identify users' true consumption intentions in different regions. Furthermore, existing technologies lack the ability to detect network proxy characteristics, resulting in delivery deviations when a user's actual location doesn't match their IP address. This solution, however, effectively eliminates interference from false geographic locations by analyzing IP hop frequency and proxy characteristics.

[0086] Through the above technical solution, this application can solve the problem of one-sided user behavior data in cross-border advertising. By integrating dynamic interactive behaviors, transaction preferences, and device network characteristics, it provides a more comprehensive data foundation for the subsequent generation of cross-border user profiles. For example, the association between language preference and transaction currency can help determine a user's multilingual ability and consumption regional tendencies, and the combination of device time zone and IP hopping characteristics can improve the accuracy of identifying the user's true geographic location.

[0087] In one possible implementation, reference Figure 3 , step S200 includes steps S210 to S240, wherein:

[0088] Step S210: De-identify the user's page dwell time and language preferences in multiple countries or regions to generate a user behavior link code containing multiple language preferences;

[0089] Step S220: Correlate and match the transaction data with the click behavior sequence to generate a user product category preference label;

[0090] Step S230 , extracting user cross-time zone activity features based on device time zone, IP address hopping frequency, and network proxy features;

[0091] In step S240, the user behavior link code, the user product category preference label, and the user's cross-time zone active features are input into a preset federated learning model to generate a cross-border user portrait.

[0092] In this embodiment, the user behavior link encoding of multilingual preferences refers to the serialized encoding of the user's language selection and stay time on pages in different countries through a hash algorithm. Specifically, the SHA-256 algorithm can be used to combine the page language code and the stay timestamp to generate an irreversible identifier, which is used to eliminate user identity sensitive information while retaining cross-regional behavior characteristics. The user's product category preference label refers to the time-series matching of the product category in the transaction record with the product display page in the click behavior. Specifically, the Apriori association rule algorithm can be used to mine the probability threshold of the transaction generated after the user clicks the product page, and generate a strongly associated product category label. The user's cross-time zone active feature refers to the calculation of the user's active period distribution through the deviation value between the device time zone and the IP address geographic location. Specifically, a sliding time window can be used to count the time period when the IP address switching frequency exceeds the set threshold as the cross-border activity hotspot period. The preset federated learning model refers to a distributed machine learning framework deployed on servers in multiple geographic regions. Specifically, a horizontal federated learning architecture can be used to realize the encrypted aggregation of user feature vectors in each region, and the shared model parameters can be updated through gradient exchange.

[0093] Specifically, the user's language preference and duration of stay on pages in multiple countries are first de-identified. For example, the duration of stay when a user selects English on a US page and the duration of stay when a user selects Japanese on a Japanese page are hashed separately to form a cross-language behavioral link sequence. The user's click timestamp on the product details page is then matched with the transaction timestamp returned by the payment gateway. For example, if a user completes a cross-border payment for a mobile phone within three minutes of browsing the electronics page, a preference label for the electronics category is generated. Furthermore, based on the difference between the device time zone and the actual geographic location of the IP address, for example, if the device is set to the Eastern Time Zone but the IP address frequently switches between the UTC+1 and UTC+3 time zones, combined with VPN usage records, it is determined that the user has activity characteristics across European time zones. Finally, the processed feature vector is input into a federated learning model. For example, homomorphic encryption technology is used to jointly train the characteristics of EU users and Southeast Asian users to generate a user profile that reflects cross-border consumption habits.

[0094] Compared with existing technologies, traditional methods only build user profiles based on browsing data from a single geographic location and are unable to identify users' true preferences in a multilingual environment. For example, users' English browsing behavior on UK pages may be misjudged as local user needs. However, this solution can effectively correlate user behavior trajectories on pages in multiple countries with real transaction data through cross-regional feature fusion. For example, it can identify that although users browse products in English, their actual purchasing preferences point to specialty products in a specific region. The analysis of user active time periods in existing technologies is usually based on the local time of the device, while this solution can accurately capture users' true active time periods during international travel or cross-border shopping by detecting IP address hopping frequency and time zone deviation.

[0095] Through the above technical solutions, this application solves the problem of insufficient portrait accuracy caused by the fragmentation of cross-border user behavior characteristics, and realizes the organic integration of multilingual behavior data and cross-time zone activity characteristics. Specifically, through de-identification processing, it balances user privacy protection and behavioral feature availability. The federated learning model is used to improve the integrity of cross-border user portraits while protecting data privacy. The generated user portraits can accurately reflect the user's consumption preferences and activity patterns in multiple countries or regions, providing reliable data support for subsequent advertising matching.

[0096] In one possible implementation, reference Figure 4 , the step S300 includes steps S310 to S330, wherein:

[0097] Step S310, obtaining exchange rate fluctuation data in real time through the central bank interface of the target region;

[0098] Step S320: Call the API interface of the General Administration of Customs of the target country to obtain the tariff rate data corresponding to the HS code;

[0099] Step S330: Obtain logistics cost data through the logistics service provider interface; the logistics cost data includes real-time freight and estimated delivery time.

[0100] In this embodiment, the exchange rate fluctuation data refers to the real-time change in the exchange rate between the target region's currency and the advertiser's local currency. Specifically, periodic data capture can be achieved by accessing the exchange rate API interface publicly available on the central bank, and used to dynamically evaluate the currency exchange gains and losses in the advertising costs. Tariff rate data refers to the import tax rate imposed by the target country on specific commodity categories. Specifically, the HS code query interface provided by the General Administration of Customs can be used to match commodity categories and tax rates, and is used to calculate the tax costs of advertised commodities when they enter the target market. Logistics cost data refers to the real-time freight and delivery time of commodities from the advertiser's warehouse to the target region. Specifically, the cost and time parameters under different transportation methods can be obtained through the logistics service provider interface, and are used to comprehensively evaluate the impact of the distribution link on advertising decisions.

[0101] Specifically, exchange rate fluctuation data is collected through the central bank interface at preset intervals, for example, updated every five minutes, to ensure real-time economic parameters for advertising. Tariff rate data is mapped to advertised product categories using HS codes. For example, clothing products correspond to HS code sections 61-63, and the corresponding tax rate values ​​are returned by calling the General Administration of Customs API. Logistics cost data is obtained by inputting product weight, volume, and destination coordinates through the logistics service provider interface, which returns the freight and estimated delivery time for currently available transportation channels. These three types of data are integrated and combined with the match between the ad and the user, using a weighted calculation to generate economic parameters for advertising, providing a basis for cost-benefit evaluation of subsequent ad generation.

[0102] Compared to existing technologies, traditional cross-border advertising typically relies solely on fixed exchange rates and historical tariff data for cost estimation, resulting in economic parameter calculations lagging behind market changes. This solution, however, leverages central bank interfaces, the General Administration of Customs API, and logistics service provider interfaces to enable real-time, simultaneous access to multi-dimensional data. This dynamically reflects the impact of exchange rate fluctuations, policy adjustments, and changes in logistics capacity on advertising. For example, if the target region's currency suddenly depreciates, real-time exchange rate data can immediately trigger a recalculation of economic parameters for advertising, avoiding cost errors caused by delayed exchange rate updates.

[0103] Through the above technical solution, this application can accurately quantify the economic factors affecting cross-border advertising, resolving the cost calculation bias caused by lagging data updates in traditional methods. By acquiring exchange rate, tariff, and logistics data in real time, the economic parameters of advertising can dynamically reflect changes in the target region's market environment, allowing the ad generation process to automatically avoid high-cost areas caused by sudden economic fluctuations. At the same time, it prioritizes promotional routes with stable logistics timeliness and low taxes, thereby improving the input-output ratio of advertising.

[0104] In a feasible implementation, the step of calculating the economic parameters of advertising based on the matching degree between advertisements and users, exchange rate fluctuation data, tariff rate data, and logistics cost data specifically includes:

[0105] The economic parameters of advertising are calculated according to the following formula:

[0106] ;

[0107] in, Indicates the economic parameters of advertising placement; Indicates the matching degree between advertisement and user; Indicates exchange rate fluctuation data; Represents tariff rate data; Indicates real-time shipping rates; Indicates the estimated delivery time; represents the exchange rate gain weight; represents the tariff loss weight; Indicates the logistics freight weight; Represents the logistics timeliness weight.

[0108] In this embodiment, the matching degree between advertisements and users refers to the degree of association between advertisement content and user preferences. Specifically, it can be achieved through the cosine similarity algorithm between the user behavior link code and the advertisement tag, and is used to reflect the attractiveness of the advertisement to the target user. Exchange rate fluctuation data refers to the real-time exchange rate change of the target region's currency against the base currency. Specifically, it can be achieved through the calculation of the exchange rate mid-price difference of the central bank interface, and is used to measure the impact of cross-border transaction cost fluctuations on advertising revenue. Tariff rate data refers to the import tax rate levied by the target country on specific commodity categories. Specifically, it can be achieved by querying the tax rate corresponding to the HS code through the General Administration of Customs API interface, and is used to evaluate the compliance cost of the goods after entry. Real-time freight refers to the logistics fee required to transport the goods from the advertiser's location to the target area. Specifically, it can be achieved by obtaining real-time quotes through the logistics service provider interface, and is used to calculate the impact of logistics costs on user purchasing intention. Estimated delivery time refers to the time period from the shipment of goods to the user's receipt. Specifically, it can be achieved by combining the historical data of the logistics service provider with the current transportation capacity status forecast, and is used to evaluate the negative effect of logistics time on advertising conversion rate. The exchange rate gain weight, tariff loss weight, logistics freight weight and logistics time efficiency weight refer to the influence coefficients of different economic factors on the economic efficiency of advertising. They can be determined through multivariate regression analysis of historical advertising data and used to dynamically adjust the proportion of each factor in the evaluation of advertising economic efficiency.

[0109] Specifically, in the process of calculating the economic parameters for advertising, the matching degree between advertising and users is first used as a basic parameter to reflect the degree of fit between advertising and user needs. Next, exchange rate fluctuation data is collected in real time through the interface of the central bank of the target region, and the gain or loss of the exchange rate change on the economic efficiency of advertising is calculated in combination with the exchange rate gain weight. At the same time, the API interface of the General Administration of Customs is called to obtain the current tariff rate data, and the tax cost of the goods after entry is evaluated based on the tariff loss weight. In addition, real-time freight and estimated delivery time are collected through the logistics service provider interface, and the negative impact of logistics costs on user decisions is quantified by combining the logistics freight weight and the logistics time weight respectively. Finally, the above parameters are substituted into the preset economic parameter calculation formula to dynamically generate the economic parameters for advertising, providing a quantitative decision-making basis for subsequent advertising generation and delivery.

[0110] Compared to existing technologies, traditional advertising economics evaluation typically relies solely on single metrics such as click-through rate or conversion rate, failing to consider economic factors unique to cross-border scenarios, such as exchange rate fluctuations, tariff rates, and logistics costs. For example, existing technologies may adjust advertising strategies based solely on user click behavior, but fail to integrate the impact of tariff rate changes on the final selling price of goods, resulting in lower-than-expected conversion rates after advertising. This solution incorporates exchange rate risks, tariff costs, and logistics timelines in cross-border transactions into the advertising economics evaluation system, enabling dynamic optimization of advertising strategies.

[0111] Through the above technical solution, this application can effectively address the problem of advertising revenue loss caused by ignoring economic fluctuations in traditional cross-border advertising. By quantifying the impact of exchange rate fluctuations on transaction costs, the restrictions of tariff rates on commodity pricing, and the negative impact of logistics time on user experience, advertising economic parameters can dynamically reflect changes in the target market's economic environment, helping advertisers avoid high-risk areas in advance when generating ads and optimize advertising resource allocation.

[0112] In one possible implementation, reference Figure 5 , the step S500 includes steps S510 to S530, wherein:

[0113] Step S510: Based on the advertising economic parameters and the product category preference tags in the cross-border user profile, a corresponding advertising template is called from a preset advertising material library, and an original advertisement is generated in combination with the advertising economic parameters;

[0114] Step S520: Parse the advertisement review rules in the target region's legal database and perform sensitive word detection and contraband content filtering on the original advertisement;

[0115] Step S530 , performing compliance verification on the visual elements of the advertisement using the taboo color library, religious symbol library, and festival custom library in the cultural library of the target region to generate a final advertisement.

[0116] In this embodiment, the advertising material library refers to a database that stores different advertising templates and materials. Specifically, it can be implemented by a distributed storage system combined with a label classification index, and is used to quickly match advertising templates according to product category preference labels. The legal library refers to a data set containing advertising laws and regulations in the target area. Specifically, it can be implemented by obtaining government public documents or accessing a third-party compliance database, and is used to detect whether the advertising content complies with local legal requirements. The cultural library refers to a knowledge base that stores information on cultural taboos in the target area. Specifically, a multimodal data collection tool can be used to integrate religious symbols, festival customs, and color taboo data to verify the compliance of advertising visual elements. Sensitive word detection refers to identifying illegal words in advertising texts. Specifically, it can be implemented by using a natural language processing model combined with regular expression matching rules to filter content that violates laws or cultural taboos. Compliance verification refers to verifying whether advertising images, colors, and symbols comply with the cultural habits of the target area. Specifically, it can be implemented by using an image recognition algorithm combined with a rule engine to avoid advertising placement conflicts due to cultural differences.

[0117] Specifically, this step first matches user preferences with ad templates through the ad creative library, combining economic parameters to generate the original ad content. The ad text is then scanned for sensitive words and prohibited items using the review rules of the legal library, detecting, for example, whether there is false advertising or information about restricted goods. The taboo color library in the cultural library is then used to verify the main color tone of the ad. Any taboo colors in the target region are automatically replaced with safe color values. The religious symbol library is also used to identify potentially conflicting elements in the ad's graphics, such as removing graphics that are inconsistent with local beliefs. Finally, the time points in the ad copy are adjusted based on the festival customs library, such as inserting relevant greetings during specific festivals. After multiple layers of filtering and format conversion, the final ad is generated, complying with the legal regulations and cultural customs of the target region.

[0118] Compared to existing technologies, existing cross-border ad generation methods typically rely solely on simple keyword matching or geolocation information, failing to consider the synergy between legal review and cultural adaptation. For example, traditional solutions might push ads based solely on IP addresses but fail to identify religious symbol conflicts or color taboos implicit in the ad content. This solution, by leveraging a dual verification mechanism based on legal and cultural databases, achieves a multi-dimensional match between ad content and the regulations of the region where it is being advertised, addressing compliance risks associated with single-dimensional filtering.

[0119] Through the above technical solution, this application can effectively avoid advertising violations caused by legal differences or cultural misunderstandings, and improve the legality and cultural adaptability of advertising content in the target region. At the same time, through automated filtering and format conversion processes, it can reduce manual review costs and improve the efficiency of cross-border advertising.

[0120] In one possible implementation, reference Figure 6, the step S600 includes steps S610 to S620, wherein:

[0121] Step S610: Adjust the format parameters of the final advertisement according to the terminal device resolution distribution data in the target area, including the video bit rate, the location of the multilingual subtitles, and the size of the interactive button;

[0122] Step S620 : converting the final advertisement into a standardized format that meets the requirements of the advertisement server cluster in the target region, and distributing the final advertisement to the target time period server nodes that meet the user's cross-time zone active characteristics.

[0123] In this embodiment, the target area terminal device resolution distribution data refers to a statistical collection of screen resolution information of user terminal devices in the target area. Specifically, it can be achieved by analyzing the screen parameters in the device feature data or calling a third-party data analysis platform interface to dynamically match the advertising display specifications to avoid interface adaptation anomalies.

[0124] Video bitrate adjustment refers to dynamically adjusting the compression rate of video ads based on the peak value of the resolution distribution. This can be achieved by using the hierarchical compression algorithm of the H.265 encoding protocol to ensure smooth playback on devices with different resolutions.

[0125] Optimizing the embedding position of multilingual subtitles means automatically adjusting the subtitle display area based on the device screen ratio. This can be achieved through the edge detection algorithm of the OpenCV image processing library to prevent subtitles from blocking key visual elements.

[0126] Interactive button size calibration refers to setting the minimum pixel threshold of touch buttons according to the resolution range. This can be achieved by using the breakpoint adaptation mechanism of the responsive layout framework to ensure user operation accuracy.

[0127] The cross-time zone user activity feature refers to the user's active time period data in multiple time zones. This can be achieved through correlation analysis between the device time zone and the timestamp of the server log, which is used to match the timeliness window for advertising delivery.

[0128] Specifically, during the ad delivery phase, device resolution statistics for the target region are first collected. For example, a resolution distribution histogram is generated by sampling screen width and height parameters from device feature data. Based on this histogram, the mainstream resolution range is determined, and dynamic bitrate adjustment technology is used to adaptively compress video ads. For example, a 1080p video may be switched to a 720p bitrate when a low-resolution device is detected. Furthermore, subtitles are offset based on the screen aspect ratio. For example, on a 16:9 screen, subtitles are placed in the bottom center, while on an 18:9 screen, they are shifted 5% of the display height. The size of interactive buttons is scaled proportionally to the physical size density of the resolution range, for example, setting a button height of at least 48 pixels on a 5-inch screen. After format adjustment, the ad content is encapsulated into a standardized transmission format, such as the HLS streaming protocol or AMP HTML format. Ads are then distributed to edge server nodes in the corresponding time zone based on the user's active time period as recorded in the user's cross-time zone activity profile. For example, if a user is active between 8:00 PM and 10:00 PM Beijing time, the ad is preloaded into the cache queue of a CDN node in the East 8 region.

[0129] Compared with existing technologies, traditional ad delivery typically uses fixed-resolution templates or a single time zone delivery strategy, resulting in cropped images on low-resolution devices or ineffective ad delivery during high-latency periods. This solution dynamically adapts formats by analyzing resolution distribution data. For example, if it detects that 40% of users in a region use 720×1280 resolution devices, it prioritizes ad versions adapted to that resolution. It also incorporates cross-time zone activity patterns to precisely schedule server nodes. For example, based on historical user behavior, ads for users in the western United States can be pushed to Los Angeles servers two hours in advance.

[0130] Through the above technical solution, this application solves the problem of misalignment of advertising elements caused by poor display adaptability of devices in different regions, avoids the waste of advertising exposure caused by time zone matching errors, and significantly improves the rendering completeness of advertising materials and the timeliness of user reach, thereby increasing effective click behavior and reducing redundant consumption of server resources.

[0131] In one possible implementation, reference Figure 7 After step S600, the method further includes steps S710 to S740, wherein:

[0132] Step S710: Collect the exposure, click-through rate, and conversion rate of the final advertisement;

[0133] Adjust the weight of the user's cross-time zone active feature of the federated learning model based on the exposure, click-through rate, and conversion rate to generate an advertising optimization coefficient;

[0134] The exposure volume is matched with the number of active users in the target area during the time period using a decay function, the click-through rate and the product category preference label are subjected to cosine similarity analysis, and the conversion rate and the economic parameters of the advertising are subjected to Pearson correlation verification to generate a multi-dimensional evaluation matrix;

[0135] The final advertisement is optimized according to the advertisement optimization coefficient and the multi-dimensional evaluation matrix.

[0136] In this embodiment, exposure refers to the number of times an advertisement is displayed in the target area. Specifically, data collection can be achieved through the advertising server log analysis tool to reflect the breadth of advertisement coverage. Click-through rate refers to the ratio of the number of times an advertisement is clicked to the amount of exposure. Specifically, tracking statistics can be achieved through front-end embedding technology to measure the attractiveness of advertising content. Conversion rate refers to the proportion of users who complete purchases or registrations after clicking. Specifically, monitoring can be achieved through payment interface callbacks or behavior tracking SDKs to evaluate the actual effectiveness of advertisements. Federated learning model weight adjustment refers to dynamically correcting the contribution of users' cross-time zone active features in the portrait based on advertising effect feedback. Specifically, a gradient descent algorithm can be used to achieve parameter updates to optimize the accuracy of user behavior predictions. The multi-dimensional evaluation matrix refers to a comprehensive evaluation system that integrates exposure attenuation matching, preference similarity analysis, and economic parameter correlation verification. Specifically, data association analysis can be achieved through a matrix operation engine to identify the optimization direction of advertising delivery strategies.

[0137] Specifically, after the ad is delivered, the log system of the ad server cluster collects exposure data in real time, and the user behavior tracking module is used to calculate the click-through rate and conversion rate. The federated learning model automatically adjusts the weight parameters of the user's cross-time zone active features based on the ad optimization coefficient. For example, when the conversion rate of a certain time period increases significantly, the weight ratio of the active features of that time period is increased accordingly. At the same time, the decay function is used to calculate the matching degree between the exposure and the number of active users in the time period to identify whether the ad delivery period overlaps with the user's actual active period; the cosine similarity is used to analyze the correlation between the click-through rate and the product category preference label to verify whether the ad content is in line with the user's interests; the Pearson correlation is used to verify the relationship between the conversion rate and the economic parameters of the ad delivery to determine whether the cost control strategy affects the conversion effect. Finally, the multi-dimensional evaluation results are input into the optimization algorithm to generate instructions for replacing advertising materials, adjusting time periods, or correcting economic parameters.

[0138] Compared to existing technologies, existing ad optimization methods typically rely solely on a single metric, such as click-through rate or conversion rate, to adjust strategies, failing to effectively identify complex issues such as time-matching mismatch, interest deviation, or cost control overload. This solution dynamically optimizes the weights of user behavior features through a federated learning model. Combining a decay function, similarity analysis, and correlation verification, this solution constructs a multi-dimensional evaluation system that simultaneously addresses issues such as misaligned ad delivery timeframes, insufficiently attractive content, and cost-benefit imbalances.

[0139] Through the above technical solution, this application can dynamically optimize advertising delivery strategies based on real-time performance data, balance economic cost constraints while ensuring user interest matching, effectively improve the accuracy and conversion efficiency of cross-border advertising, and avoid resource waste or decreased user experience due to single indicator optimization.

[0140] It should be noted that the above examples are only used to understand this application and do not constitute a limitation on the cross-border advertising precision delivery method based on user portraits in this application. More simple transformations based on this technical concept are all within the scope of protection of this application.

[0141] This application also provides a cross-border advertising precision delivery system 10 based on user portraits, Figure 8 , the system comprising:

[0142] The first data acquisition module 100 is used to acquire user behavior data, transaction data, and device feature data;

[0143] A cross-border user profile generation module 200 is used to perform cross-regional feature fusion processing on the behavior data, transaction data, and device feature data to generate a cross-border user profile;

[0144] The second data acquisition module 300 is used to determine the matching degree between the advertisement and the user based on the cross-border user profile, and obtain exchange rate fluctuation data, tariff rate data and logistics cost data of the target region;

[0145] Advertisement placement economic parameter calculation module 400, for calculating advertisement placement economic parameters based on the matching degree between advertisements and users, exchange rate fluctuation data, tariff rate data, and logistics cost data;

[0146] The final advertisement generation module 500 is used to generate an original advertisement based on the cross-border user profile and the economic parameters of the advertisement placement, and to filter and convert the original advertisement according to the legal and cultural databases of the target region to generate the final advertisement;

[0147] The advertisement delivery module 600 is configured to match the final advertisement to an advertisement server cluster in a target area for advertisement delivery.

[0148] In this embodiment, the first data acquisition module 100 refers to a device that collects user multi-regional behaviors and transaction records through a browser interface and a payment gateway. Specifically, it can be implemented by a joint calling mechanism of a browser API and a payment gateway interface to eliminate the regional coverage deviation caused by a single data source. The cross-border user portrait generation module 200 refers to a processor 30 that integrates cross-regional behavior links and transaction preferences. Specifically, it can be implemented by a multimodal feature fusion algorithm under a federated learning framework to solve the problem of user behavior fragmentation in cross-border scenarios. The advertising economic parameter calculation module 400 refers to an operation unit that dynamically integrates exchange rate fluctuations and logistics costs. Specifically, it can be implemented by a dynamic adjustment algorithm of weighted economic indicators to avoid hidden cost risks in cross-border transactions. The final advertising generation module 500 refers to an advertising compliance device that combines legal rules and cultural taboos. Specifically, it can be implemented by multi-level semantic filtering and visual element replacement technology to meet the content review requirements of different legal domains.

[0149] Specifically, the system achieves precise cross-border advertising through a collaborative processing mechanism of multi-source heterogeneous data. The first data acquisition module 100 collects multi-dimensional cross-border behavioral characteristics of users from browsers, payment systems and terminal devices. The cross-border user portrait generation module 200 uses a federated learning model to convert discrete cross-regional behavior codes into a unified portrait. The second data acquisition module 300 synchronously accesses the central bank and customs data interfaces to obtain the economic fluctuation parameters of the target area in real time. The advertising delivery economic parameter calculation module 400 couples user preferences with economic variables through a dynamic weight distribution model to generate delivery parameters that include cost-effectiveness forecasts. The final advertising generation module 500 performs dual compliance transformation on advertising content based on the text review rules of the legal library and the visual verification rules of the cultural library. The advertising delivery module 600 implements automatic adaptation and distribution of advertising formats based on the device characteristics and network environment of the target area.

[0150] Compared to existing technologies, traditional cross-border advertising systems rely solely on static geographic location information for delivery. This system, by integrating cross-regional features with dynamic economic parameters, establishes a multi-dimensional model linking user preferences and the economic environment. Existing technologies lack automated adaptation mechanisms for legal and cultural differences. This system, through a structured legal rules library and a cultural taboo library, enables real-time compliance verification of advertising content. Existing systems utilize a fixed-cost accounting model. This system improves the accuracy of economic benefit forecasts for advertising through a dynamic weighting algorithm that combines exchange rate fluctuation data with logistics costs.

[0151] Through the above technical solutions, this application effectively solves the three major technical bottlenecks in cross-border advertising: incomplete user portraits, lagging economic parameters, and legal and cultural conflicts. The cross-regional feature fusion mechanism can accurately identify users' cross-border consumption preferences, the dynamic economic parameter model can reflect the hidden cost changes of the target market in real time, and the multi-level compliance verification system can avoid advertising failures caused by regional and cultural differences. This system significantly improves the accuracy and compliance of cross-border advertising, reduces the economic risks caused by exchange rate fluctuations or tariff adjustments, and reduces manual adaptation costs through automated format conversion.

[0152] This application also provides a cross-border advertising precision delivery device based on user portraits, refer to Figure 9 The device includes: a memory 20, a processor 30, and a user portrait-based cross-border advertising precision delivery program stored on the memory 20 and runnable on the processor 30. The user portrait-based cross-border advertising precision delivery program is configured to implement the steps of the user portrait-based cross-border advertising precision delivery method.

[0153] The cross-border advertising precision delivery device based on user portraits provided in this application adopts the cross-border advertising precision delivery method based on user portraits in the above-mentioned embodiment, which can improve the accuracy of cross-border advertising delivery, conversion effect and reduce compliance risks. Compared with the existing technology, the beneficial effects of the cross-border advertising precision delivery device based on user portraits provided in this application are the same as the beneficial effects of the cross-border advertising precision delivery method based on user portraits provided in the above-mentioned embodiment, and the other technical features of the cross-border advertising precision delivery device based on user portraits are the same as the features disclosed in the above-mentioned embodiment method, which will not be repeated here.

[0154] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for precise cross-border advertising delivery based on user portraits, characterized in that: The method includes: Obtain user behavior data, transaction data, and device feature data; Perform cross-regional feature fusion processing on the behavioral data, transaction data, and device feature data to generate cross-border user profiles; Determine the matching degree between the advertisement and the user based on the cross-border user profile, and obtain exchange rate fluctuation data, tariff rate data, and logistics cost data for the target region; Calculate advertising economic parameters based on the matching degree between advertisements and users, exchange rate fluctuation data, tariff rate data, and logistics cost data; Generate original ads based on cross-border user profiles and advertising economic parameters. Filter and convert the original ads according to the legal and cultural databases of the target region to generate the final ads. The final advertisement is matched to an advertisement server cluster in a target area for advertisement delivery.

2. The cross-border advertising precision delivery method based on user portraits according to claim 1 is characterized in that: The steps of obtaining user behavior data, transaction data and device feature data include: Collecting the behavioral data through the browser API; the behavioral data includes the user's page dwell time, click behavior sequence and language preference in multiple countries or regions; Obtain the transaction data through the payment gateway interface; the transaction data includes transaction currency, transaction amount and commodity category; The device characteristic data is obtained through the terminal device used by the user; the device characteristic data includes the device time zone, IP address hopping frequency and network proxy characteristics.

3. The cross-border advertising precision delivery method based on user portraits according to claim 2 is characterized in that: The step of performing cross-regional feature fusion processing on the behavior data, transaction data, and device feature data to generate a cross-border user profile includes: De-identifying the user's page dwell time and language preferences in multiple countries or regions to generate a user behavior link code containing multiple language preferences; Correlating and matching the transaction data with the click behavior sequence to generate a user product category preference label; Extract user activity characteristics across time zones based on device time zones, IP address hopping frequency, and network proxy features; The user behavior link code, user product category preference label, and user cross-time zone active features are input into a preset federated learning model to generate a cross-border user profile.

4. The cross-border advertising precision delivery method based on user portraits according to claim 1 is characterized in that: The step of obtaining the exchange rate data, tariff data and logistics cost data of the target region and calculating the economic parameters of the advertising based on the exchange rate data, tariff data and logistics cost data includes: Obtain real-time exchange rate fluctuation data through the central bank interface of the target region; Call the target country's General Administration of Customs API to obtain the tariff rate data corresponding to the HS code; Obtain logistics cost data through the logistics service provider interface; the logistics cost data includes real-time freight and estimated delivery time.

5. The cross-border advertising precision delivery method based on user portraits according to claim 4 is characterized in that: The step of calculating the economic parameters of advertising based on the matching degree between advertisements and users, exchange rate fluctuation data, tariff rate data, and logistics cost data specifically includes: The economic parameters of advertising are calculated according to the following formula: ； in, Indicates the economic parameters of advertising placement; Indicates the matching degree between advertisement and user; Indicates exchange rate fluctuation data; Represents tariff rate data; Indicates real-time shipping rates; Indicates the estimated delivery time; represents the exchange rate gain weight; represents the tariff loss weight; Indicates the logistics freight weight; Represents the logistics timeliness weight.

6. The cross-border advertising precision delivery method based on user portraits according to claim 1 is characterized in that: The steps of generating the original advertisement based on the cross-border user profile and the economic parameters of the advertising placement, filtering and formatting the original advertisement according to the legal and cultural databases of the target region, and generating the final advertisement include: Based on the economic parameters of advertising and the product category preference tags in the cross-border user profile, the corresponding advertising template is called from the preset advertising material library, and the original advertisement is generated in combination with the economic parameters of advertising. Parse the advertising review rules in the target region's legal database and perform sensitive word detection and prohibited content filtering on original ads; The compliance of the visual elements of the advertisement is checked against the taboo color library, religious symbol library and festival customs library in the cultural library of the target area to generate the final advertisement.

7. The cross-border advertising precision delivery method based on user portraits according to claim 1 is characterized in that: The step of matching the final advertisement to an advertisement server cluster in a target area for advertisement delivery includes: Adjust the final ad format parameters based on the target region's device resolution distribution data, including video bitrate, multilingual subtitle embedding location, and interactive button size; The final advertisement is converted into a standardized format that meets the requirements of the ad server cluster in the target region, and distributed to the server nodes in the target time period that meets the user's active characteristics across time zones.

8. The cross-border advertising precision delivery method based on user portraits according to claim 1 is characterized in that: After the step of matching the final advertisement to an advertisement server cluster in a target area for advertisement delivery, the method further includes: Collect the final ad impressions, click-through rates, and conversion rates; Adjust the weight of the user's cross-time zone active feature of the federated learning model based on the exposure, click-through rate, and conversion rate to generate an advertising optimization coefficient; The exposure volume is matched with the number of active users in the target area during the time period using a decay function, the click-through rate and the product category preference label are subjected to cosine similarity analysis, and the conversion rate and the economic parameters of the advertising are subjected to Pearson correlation verification to generate a multi-dimensional evaluation matrix; The final advertisement is optimized according to the advertisement optimization coefficient and the multi-dimensional evaluation matrix.

9. A cross-border advertising precision delivery system based on user portraits, characterized by: The system comprises: A first data acquisition module is used to acquire user behavior data, transaction data and device feature data; A cross-border user profile generation module is used to perform cross-regional feature fusion processing on the behavioral data, transaction data, and device feature data to generate a cross-border user profile; A second data acquisition module is used to determine the matching degree between the advertisement and the user based on the cross-border user profile, and to obtain exchange rate fluctuation data, tariff rate data, and logistics cost data of the target region; Advertisement delivery economic parameter calculation module, used to calculate the economic parameters of advertisement delivery based on the matching degree between advertisement and user, exchange rate fluctuation data, tariff rate data and logistics cost data; The final ad generation module is used to generate original ads based on cross-border user profiles and economic parameters of advertising placement, and to filter and format the original ads according to the legal and cultural databases of the target region to generate the final ads; The advertisement delivery module is used to match the final advertisement to the advertisement server cluster in the target area for advertisement delivery.

10. A device for precise cross-border advertising delivery based on user portraits, characterized in that: The device includes: a memory, a processor, and a user portrait-based cross-border advertising precision delivery program stored in the memory and runnable on the processor, wherein the user portrait-based cross-border advertising precision delivery program is configured to implement the steps of the user portrait-based cross-border advertising precision delivery method as described in any one of claims 1 to 8.

Images