Personalized advertisement putting method and system based on big data

Abstract

The invention discloses a personalized advertisement putting method and system based on big data, and the method comprises the steps: obtaining user behavior data, equipment information, geographic position information, context data and third-party data from a plurality of sources, and storing the data in different formats in a data platform in a unified manner through data processing; constructing a multi-dimensional dynamic user portrait, and performing real-time or near-real-time updating according to the real-time behavior data of the user; dynamic advertisement content generation and selection based on user tolerance: dynamically generating advertisement texts, images, videos and interactive elements, analyzing the tolerance of users to different types of advertisements in combination with user historical feedback and real-time behavior data, and selecting optimal advertisement content; and based on the user portrait and the advertisement content, carrying out real-time advertisement putting decision making by adopting reinforcement learning and a dobby machine algorithm, and dynamically adjusting a putting strategy according to user clicking and browsing duration and conversion real-time feedback. According to the invention, the accuracy and economic benefits of advertisement putting can be improved while the user experience is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of big data processing, and in particular to a method and system for personalized advertising based on big data. Background Art

[0002] With the widespread adoption of the internet, mobile communications, and smart devices, digital advertising has become a vital tool in commercial marketing. Traditional advertising methods, primarily based on crude demographics and static user profiles, lack the ability to sensitively capture the dynamic changes in individual user interests. This leads to low ad hit rates, poor user experience, and insufficient return on investment (ROI).

[0003] In recent years, the development of big data technology has made it possible to collect, store, and process massive amounts of user data. By analyzing users' browsing behavior, interests, geographic location, device information, social interactions, and third-party data, we can build more refined and dynamic user profiles, providing a foundation for personalized advertising. However, existing technologies still have the following shortcomings in practical applications: On the one hand, existing personalized advertising systems are mostly based on static or infrequently updated user profiles, which fail to reflect changes in user interests in real time. This leads to ads being misaligned with users' current needs, impacting click-through and conversion rates. On the other hand, ad content generation often relies on preset templates, lacking personalized adaptation to different user tolerances, and can easily lead to user dissatisfaction or even negative feedback.

[0004] In addition, in the advertising decision-making process, traditional methods usually adopt simple rule-based or static model-based material selection strategies, lack effective online learning mechanisms, and find it difficult to strike a balance between exploring new advertising materials and utilizing existing efficient materials, thus limiting the improvement of the overall system revenue.

[0005] Therefore, there is an urgent need for a personalized advertising delivery method based on big data that can dynamically update user portraits in real time, dynamically generate advertising content based on user tolerance, and achieve real-time decision optimization through reinforcement learning and multi-armed bandit algorithms to improve the accuracy of advertising delivery, user experience and overall conversion effect. Summary of the Invention

[0006] In order to solve the above technical problems in the prior art, the present invention proposes a method and system for personalized advertising based on big data to solve the above technical problems.

[0007] According to a first aspect of the present invention, a method for personalized advertising based on big data is proposed, comprising: S1: Obtain user behavior data, device information, geographic location information, contextual data, and third-party data from multiple sources, and process and store data in different formats in a unified data platform; S2: Build multi-dimensional dynamic user profiles and update them in real time or near real time based on real-time user behavior data; S3: Dynamic ad content generation and selection based on user tolerance. Ad text, images, videos, and interactive elements are dynamically generated. The system analyzes user tolerance for different ad types based on historical user feedback and real-time behavioral data to select the optimal ad content. S4: Based on the user profile of S2 and the advertising content of S3, reinforcement learning and multi-armed bandit algorithms are used to make real-time advertising delivery decisions, and the delivery strategy is dynamically adjusted based on real-time feedback from user clicks, browsing time, and conversions.

[0008] In some specific embodiments, methods for obtaining data from multiple sources in S1 include Web / App SDK tracking, API interface calls, log file analysis, and crawler crawling.

[0009] In some specific embodiments, the data processing in S1 includes: data cleaning, conversion and integration, and data cleaning includes missing value processing, outlier detection and elimination, and duplicate value filtering.

[0010] In some specific embodiments, the user portrait construction of S2 includes: extracting key features of user preference labels, consumption capacity indicators, and behavior pattern sequences; building a model based on one or a combination of regularized labeling, cluster analysis, supervised / unsupervised machine learning, and deep neural networks; based on event-driven or timed batches, incrementally or in real time updating the features of the user portrait and applying a decay mechanism to process the weight of historical data.

[0011] In some specific embodiments, dynamic advertising content generation in S3 includes: using a large language model to generate personalized copy based on user portraits and advertising goals; using a generative adversarial network or a variational autoencoder to generate advertising materials that meet user preferences; combining templated video clips with user interest tags to generate short video ads; dynamically generating interactive components based on the user's historical interaction habits; training a classification or regression model based on historical click, skip, block, and complaint data to predict the user's tolerance score for different content, and using this to adjust the style of generated content.

[0012] In some specific embodiments, S4's real-time advertising delivery decisions include: training a policy network based on long-term user value and immediate feedback; achieving a balance between new ads and delivering ads with known good results through a multi-armed bandit algorithm; and adjusting bids and material rotation frequency based on real-time click-through rate, browsing time, and conversion data.

[0013] In some specific embodiments, the multi-armed bandit algorithm specifically includes: for each target user, based on their portrait and real-time context, generating or selecting a group of candidate advertising materials, and each candidate advertisement is regarded as a bandit arm; defining reward indicators for each candidate advertisement, including user click-through rate, browsing dwell time, advertisement conversion rate, and user depth of interaction; constructing a probability model for each advertisement, and sampling and selecting advertisement arms according to the posterior distribution; after the advertisement is launched, collecting feedback data based on user behavior, and updating the expected reward estimate of each advertisement arm in real time; gradually converging the advertising material combination to improve the overall click-through rate and conversion rate of the advertisement, and continuously adaptively adjusting according to new changes in user behavior.

[0014] In some specific embodiments, it also includes: real-time monitoring of advertising delivery effect data, including display volume, click-through rate, conversion rate and user interaction, using attribution analysis and behavioral analysis to generate evaluation results, and feeding back the evaluation results to S2 to S4 to continuously optimize the delivery effect.

[0015] According to a second aspect of the present invention, a computer-readable storage medium is provided, on which one or more computer programs are stored. When the one or more computer programs are executed by a computer processor, the above-mentioned method is implemented.

[0016] According to a third aspect of the present invention, a personalized advertising delivery system based on big data is proposed, comprising: Multi-source heterogeneous data collection and integration unit: This unit is configured to obtain user behavior data, user profile data, device information, geographic location information, contextual data, and third-party data from multiple sources, and stores data in different formats in a unified manner on the data platform through data processing; Deep user profile building and updating unit: configured to build multi-dimensional dynamic user profiles and update them in real time or near real time based on real-time user behavior data; Dynamic advertising content generation and selection unit: This unit is configured to generate and select dynamic advertising content based on user tolerance. It dynamically generates advertising text, images, videos, and interactive elements, analyzes user tolerance for different types of advertising based on historical user feedback and real-time behavioral data, and selects the optimal advertising content. Real-time intelligent advertising delivery and optimization unit: This unit is configured to make real-time advertising delivery decisions based on user profiles and advertising content, using reinforcement learning and multi-armed bandit algorithms, and dynamically adjust delivery strategies based on real-time feedback from user clicks, browsing time, and conversions.

[0017] The present invention proposes a method and system for personalized advertising based on big data, which has the following technical effects: Improve the precision and dynamism of user portraits: By integrating multi-source heterogeneous data and adopting event-driven or timed batch update mechanisms, user portraits can be updated in real time or near real time, accurately capturing changes in user interests and behaviors, significantly improving the accuracy and timeliness of user portraits.

[0018] Achieve personalized dynamic generation and adaptation of advertising content: Combining technologies such as large models and generative adversarial networks, dynamically generate advertising text, images, videos, and interactive elements that meet user preferences and tolerances, effectively improving the matching degree between advertising content and user needs, and enhancing the attractiveness and click-through rate of advertisements.

[0019] Optimizing advertising placement decisions and effectiveness: By introducing reinforcement learning strategies and multi-armed bandit algorithms, the present invention can dynamically balance the exploration of new advertising materials and the use of efficient materials, achieve real-time adaptive advertising placement decisions, and continuously optimize placement strategies based on real-time feedback such as user clicks, browsing time, and conversions, thereby significantly improving advertising conversion rates and return on investment (ROI). BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated into and constitute a part of this specification. The accompanying drawings illustrate the embodiments and, together with the description, serve to explain the principles of the invention. Other embodiments and many of the intended advantages of the embodiments will be readily apparent as they become better understood by reference to the following detailed description. Other features, objects, and advantages of the present application will become more apparent upon reading the detailed description of the non-limiting embodiments made with reference to the following drawings: Figure 1 is an exemplary system architecture diagram to which the present application may be applied; Figure 2 This is a flowchart of a method for personalized advertising based on big data according to an embodiment of the present application; Figure 3 This is a framework diagram of a personalized advertising delivery system based on big data according to an embodiment of the present application; Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the embodiment of the present application. DETAILED DESCRIPTION

[0021] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the relevant invention are shown in the accompanying drawings.

[0022] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0023] Figure 1 An exemplary system architecture 100 is shown to which a method for personalized advertising based on big data can be applied according to an embodiment of the present application.

[0024] like Figure 1 As shown, system architecture 100 may include a data server 101, a network 102, and a host server 103. Network 102 is used to provide a medium for a communication link between data server 101 and host server 103. Network 102 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0025] The main server 103 may be a server that provides various services, such as a data processing server that processes information uploaded by the data server 101 .

[0026] It should be noted that the personalized advertising delivery method based on big data provided in the embodiment of the present application is generally executed by the main server 103. Accordingly, a personalized advertising delivery system based on big data is generally set in the main server 103.

[0027] It should be noted that the data server and master server can be either hardware or software. If implemented as hardware, they can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. If implemented as software, they can be implemented as multiple software programs or software modules (e.g., software or software modules used to provide distributed services), or as a single software program or software module.

[0028] It should be understood that Figure 1 The numbers of data servers, networks and main servers in the embodiment are merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0029] Figure 2 The following is a flow chart of a method for delivering personalized advertisements based on big data according to an embodiment of the present application. Figure 2 As shown, the method includes the following steps: S1: Obtain user behavior data, device information, geographic location information, contextual data, and third-party data from multiple sources, and store data in different formats in a unified data platform through data processing.

[0030] In specific embodiments, data collected from multiple sources can be further categorized as follows: first-party data: advertising platform registration information, browsing history, purchase history, and app usage data; second-party data: user interest tags and demographic information shared with partners or data providers; and third-party data: social media activity, weather information, and geolocation data from public or commercial channels. Data acquisition methods include web / app SDK tracking, API calls, log file analysis, and crawling.

[0031] In specific embodiments, data processing includes data cleaning, transformation, and integration. Data cleaning includes missing value processing, outlier detection and removal, and duplicate value filtering. Data cleaning involves identifying and processing missing values, outliers, and duplicates. Data transformation and integration involves converting data from logs, API pulls, and crawlers (which crawl public internet data, subject to compliance with relevant regulations and platform agreements) into a unified structured format. The processed data is then written to a data lake based on Hadoop, Spark, or Flink through batch or stream processing.

[0032] S2: Build multi-dimensional dynamic user profiles and update them in real time or near real time based on users’ real-time behavior data.

[0033] In a specific embodiment, user profile construction includes: extracting key features of user preference labels, consumption capacity indicators, and behavioral pattern sequences; building a model based on one or a combination of regularized labeling, cluster analysis, supervised / unsupervised machine learning, and deep neural networks; incrementally or in real time updating the features of the user profile based on event-driven or timed batches, and applying a decay mechanism to process historical data weights. Based on regularized labeling: labeling is performed based on predefined rules and logic according to user behavior data and attribute information; cluster analysis: using statistical methods such as cluster analysis and association rule mining to discover user groups and associations between users; supervised learning uses users' historical behavior data and known label information to train classification or regression models to predict users' attributes and preferences; unsupervised learning: using clustering algorithms (e.g., K-Means, DBSCAN) to group users and discover potential user groups; deep learning: using neural network models (e.g., RNN, Transformer) to process sequence data (e.g., browsing history, purchase records) to capture long-term dependencies and complex patterns of user behavior.

[0034] In a specific embodiment, the attenuation mechanism for processing historical data weights can be specifically: introducing a time attenuation function, for example, for user behavior feature values Introduce the formula for weight decay over time: ,in: Indicates a user behavior characteristic value (such as click frequency); Indicates the time difference between the behavior and the current time; Represents the decay coefficient, which adjusts the degree of forgetting of historical behavior. This setting is used to emphasize the importance of fresh behavior when dynamically updating user profile weights.

[0035] S3: Dynamic advertising content generation and selection based on user tolerance. Ad text, images, videos, and interactive elements are dynamically generated. The user's tolerance for different types of ads is analyzed based on historical user feedback and real-time behavioral data to select the optimal ad content.

[0036] In a specific embodiment, a large language model is used to generate personalized copy based on user portraits and advertising goals; a generative adversarial network or a variational autoencoder is used to generate advertising materials that meet user preferences; templated video clips are combined with user interest tags to generate short video ads; interactive components are dynamically generated based on the user's historical interaction habits; a classification or regression model is trained based on historical click, skip, block and complaint data to predict the user's tolerance score for different content, and to adjust the generated content style accordingly.

[0037] In a specific embodiment, the tolerance score can be obtained by training a regression model Given, formalized as: , where TAS stands for Tolerable Ad Score; Represents features such as ad category, length, emotional color, user behavior feedback, etc. represents the regression weight; represents the bias term.

[0038] S4: Based on the user profile of S2 and the advertising content of S3, reinforcement learning and multi-armed bandit algorithms are used to make real-time advertising delivery decisions, and the delivery strategy is dynamically adjusted based on real-time feedback from user clicks, browsing time, and conversions.

[0039] In a specific embodiment, a policy network is trained based on long-term user value and immediate feedback; a multi-armed bandit algorithm is used to strike a balance between new ads and ads with known good results; and bids and creative rotation frequency are adjusted based on real-time click-through rate, browsing time, and conversion data.

[0040] In a specific embodiment, the multi-armed bandit algorithm specifically includes: for each target user, based on their portrait and real-time context, generating or selecting a group of candidate advertising materials, each candidate advertisement is regarded as a bandit arm; defining reward indicators for each candidate advertisement, the reward indicators include at least one of user click-through rate, browsing dwell time, advertisement conversion rate, and user depth of interaction; building a probability model for each advertisement, and sampling and selecting advertisement arms according to the posterior distribution; after the advertisement is released, collecting feedback data based on user behavior, and updating the expected reward estimate of each advertisement arm in real time; gradually converging the advertising material combination to improve the overall click-through rate and conversion rate of the advertisement, and continuously adaptively adjusting according to new changes in user behavior.

[0041] In some other examples, in addition to the above-mentioned Thompson sampling method (building a probability model for each ad and sampling the ad arm according to the posterior distribution), the balancing strategy can also use the ε-greedy algorithm (randomly selecting any new ad with probability ε, and selecting the ad with the highest current return and known good performance with probability 1-ε) or the upper confidence bound UCB algorithm (selecting the ad with the highest current expected return and greater uncertainty, that is, weighing the mean and confidence interval).

[0042] In a specific embodiment, it also includes: real-time monitoring of advertising delivery effect data, including display volume, click-through rate, conversion rate and user interaction, using attribution analysis and behavioral analysis to generate evaluation results, and feeding back the evaluation results to S2 to S4 to continuously optimize the delivery effect.

[0043] The present invention provides a personalized advertising delivery method based on big data. By integrating and processing multi-source heterogeneous data, a real-time and dynamically updated multi-dimensional user portrait is constructed. Advertising content is dynamically generated and selected in combination with user tolerance analysis. Reinforcement learning and multi-armed bandit algorithms are used to achieve real-time advertising delivery decisions. Strategies are dynamically adjusted based on user feedback. This method can significantly improve the accuracy of advertising matching, click-through rate, and conversion rate, while enhancing the system's adaptability and continuous optimization capabilities. The method is suitable for a variety of online and offline application scenarios and has high practical value and promotion prospects.

[0044] Figure 3A framework diagram of a personalized advertising delivery system based on big data of a specific embodiment of the present application is shown, which includes a multi-source heterogeneous data collection and integration unit 301, a deep user portrait construction and update unit 302, a dynamic advertising content generation and selection unit 303 and a real-time intelligent advertising delivery and optimization unit 304. The multi-source heterogeneous data collection and integration unit 301 is configured to obtain user behavior data, user portrait data, device information, geographic location information, context data and third-party data from multiple sources, and uniformly store data of different formats on the data platform through data processing; the deep user portrait construction and update unit 302 is configured to construct a multi-dimensional dynamic user portrait, and update it in real time or near real time based on the user's real-time behavior data; the dynamic advertising content generation and selection unit 303 is configured to generate and select dynamic advertising content based on user tolerance, dynamically generate advertising text, images, videos and interactive elements, and analyze the user's tolerance for different types of advertisements in combination with user historical feedback and real-time behavior data to select the optimal advertising content; the real-time intelligent advertising delivery and optimization unit 304 is configured to make real-time advertising delivery decisions based on user portraits and advertising content, using reinforcement learning and multi-armed bandit algorithms, and dynamically adjust the delivery strategy based on real-time feedback on user clicks, browsing time and conversions. Each unit of the system can execute the aforementioned Figure 2 The specific steps of the method.

[0045] Reference below Figure 4 , which shows a structural diagram of a computer system suitable for implementing an electronic device of an embodiment of the present application. Figure 4 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0046] like Figure 4 As shown, the computer system includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage unit 408 into a random access memory (RAM) 403. Various programs and data required for the operation of the system 400 are also stored in the RAM 403. The CPU 401, ROM 402, and RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0047] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, a mouse, and the like; an output section 407 including devices such as a liquid crystal display (LCD) and speakers; a storage section 408 including devices such as a hard disk; and a communication section 409 including a network interface card such as a LAN card or a modem. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. Removable media 411, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 410 as needed, so that computer programs read from the media can be installed in the storage section 408 as needed.

[0048] In particular, according to embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for executing the method illustrated in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via communication portion 409 and / or installed from removable media 411. When executed by central processing unit (CPU) 401, the computer program performs the functions defined in the methods of this application. It should be noted that the computer-readable storage medium of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In this application, a computer-readable storage medium may be any tangible medium that contains or stores a program for use by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such a propagated data signal may take various forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination of the foregoing. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium that can transmit, propagate, or transfer a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0049] Computer program code for performing the operations of the present application can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0050] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0051] The modules described in the embodiments of the present application may be implemented in software or hardware.

[0052] As another aspect, the present application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or it may exist independently and not be assembled into the electronic device. The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the electronic device, the electronic device: obtains user behavior data, device information, geographic location information, context data and third-party data from multiple sources, and stores data of different formats in a data platform through data processing; constructs a multi-dimensional dynamic user portrait, and updates it in real time or near real time based on the user's real-time behavior data; generates and selects dynamic advertising content based on user tolerance, dynamically generates advertising text, images, videos and interactive elements, and analyzes the user's tolerance for different types of advertisements in combination with user historical feedback and real-time behavior data to select the optimal advertising content; based on user portraits and advertising content, uses reinforcement learning and multi-armed bandit algorithms to make real-time advertising delivery decisions, and dynamically adjusts the delivery strategy based on user clicks, browsing time and conversion real-time feedback.

[0053] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention herein is not limited to technical solutions formed by a specific combination of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A personalized advertising method based on big data, characterized in that: include: S1: Obtain user behavior data, device information, geographic location information, contextual data, and third-party data from multiple sources, and process and store data in different formats in a unified data platform; S2: Build a multi-dimensional dynamic user profile and update it in real time or near real time based on the user's real-time behavior data; S3: Dynamic ad content generation and selection based on user tolerance: Ad text, images, videos, and interactive elements are dynamically generated. The user's tolerance for different types of ads is analyzed based on historical user feedback and the real-time behavioral data to select the optimal ad content. S4: Based on the user profile of S2 and the advertising content of S3, reinforcement learning and multi-armed bandit algorithms are used to make real-time advertising delivery decisions, and the delivery strategy is dynamically adjusted based on real-time feedback from user clicks, browsing time, and conversions.

2. A method for delivering personalized advertisements based on big data according to claim 1, characterized in that: The methods of obtaining data from multiple sources in S1 include Web / App SDK tracking, API interface calls, log file analysis and crawler capture.

3. The method for delivering personalized advertisements based on big data according to claim 1, characterized in that: The data processing in S1 includes: data cleaning, conversion and integration, and the data cleaning includes missing value processing, outlier detection and elimination, and duplicate value filtering.

4. The method for delivering personalized advertisements based on big data according to claim 1, characterized in that: The user portrait construction of S2 includes: extracting key features of user preference labels, consumption capacity indicators, and behavior pattern sequences; building a model based on one or a combination of regularized labeling, cluster analysis, supervised / unsupervised machine learning, and deep neural networks; based on event-driven or timed batches, incrementally or in real time updating the features of the user portrait and applying an attenuation mechanism to process historical data weights.

5. The method for delivering personalized advertisements based on big data according to claim 1, characterized in that: The dynamic advertising content generation in S3 includes: using a large language model to generate personalized copy based on the user portrait and advertising goals; using a generative adversarial network or a variational autoencoder to generate advertising materials that meet user preferences; combining templated video clips with user interest tags to generate short video ads; dynamically generating interactive components based on the user's historical interaction habits; training a classification or regression model based on historical click, skip, block and complaint data to predict the user's tolerance score for different content, and adjusting the generated content style accordingly.

6. The method for delivering personalized advertisements based on big data according to claim 1, characterized in that: S4's real-time ad placement decisions include: training a policy network based on long-term user value and immediate feedback; achieving a balance between new ads and ads with known good results through a multi-armed bandit algorithm; and adjusting bids and creative rotation frequency based on real-time click-through rate, browsing time, and conversion data.

7. A method for delivering personalized advertisements based on big data according to claim 6, characterized in that: The multi-armed bandit algorithm specifically includes: for each target user, based on their portrait and real-time context, generating or selecting a group of candidate advertising materials, and each candidate advertisement is regarded as a bandit arm; defining reward indicators for each candidate advertisement, including user click-through rate, browsing dwell time, advertisement conversion rate, and user depth of interaction; constructing a probability model for each advertisement, and sampling and selecting advertisement arms according to the posterior distribution; after the advertisement is released, feedback data is collected based on the user behavior, and the expected reward estimate of each advertisement arm is updated in real time; gradually converging the advertising material combination to improve the overall click-through rate and conversion rate of the advertisement, while continuously adaptively adjusting according to new changes in user behavior.

8. The method for delivering personalized advertisements based on big data according to claim 1, characterized in that: Also includes: Monitor advertising effectiveness data in real time, including display volume, click-through rate, conversion rate, and user interaction, generate evaluation results using attribution analysis and behavioral analysis, and feed the evaluation results back to S2 to S4 to continuously optimize the delivery effect.

9. A computer-readable storage medium having one or more computer programs stored thereon, characterized in that: When the one or more computer programs are executed by a computer processor, the method according to any one of claims 1 to 8 is implemented.

10. A personalized advertising delivery system based on big data, characterized in that: include: Multi-source heterogeneous data collection and integration unit: This unit is configured to obtain user behavior data, device information, geographic location information, contextual data, and third-party data from multiple sources, and stores data in different formats in a unified data platform through data processing; Deep user profile building and updating unit: configured to build a multi-dimensional dynamic user profile and update it in real time or near real time based on the real-time behavior data of the user; Dynamic advertising content generation and selection unit: configured to generate and select dynamic advertising content based on user tolerance, dynamically generate advertising text, images, videos, and interactive elements, and analyze user tolerance for different types of advertising in combination with user historical feedback and the real-time behavior data to select the optimal advertising content; Real-time intelligent advertising delivery and optimization unit: configured to make real-time advertising delivery decisions based on the user portrait and the advertising content, using reinforcement learning and multi-armed bandit algorithms, and dynamically adjust the delivery strategy according to real-time feedback from user clicks, browsing time and conversions.

Images