Advertisement putting effect monitoring and optimizing system based on big data

Abstract

The invention discloses an advertisement putting effect monitoring and optimizing system based on big data, and relates to the technical field of advertisement putting, and the system comprises the following steps: carrying out the feature extraction according to advertisement putting log data, and obtaining an interaction feature set; performing user behavior prediction according to the interaction feature set to obtain an actual conversion probability; performing advertisement conversion value calculation according to the actual conversion probability to obtain an advertisement unit value quantity; performing advertisement priority ranking according to the advertisement unit value quantity to obtain a to-be-put advertisement sequence; performing budget allocation calculation according to the to-be-put advertisement sequence to obtain an advertisement budget allocation value; and performing display opportunity distribution according to the advertisement budget distribution value to obtain a display distribution result. According to the method, the use efficiency of the budget is improved, the flexibility and adaptability of budget allocation are enhanced, the risk is reduced, and the diversity and fairness of the advertisement are improved.

Description

Technical Field

[0001] The present invention relates to the field of advertising delivery technology, and in particular to an advertising delivery effect monitoring and optimization system based on big data. Background Art

[0002] With the widespread adoption of the internet and mobile devices, digital advertising has become a crucial means for businesses to gain user attention and realize commercial value. Accurately monitoring and optimizing advertising effectiveness has become a key step in improving advertising return on investment (ROI).

[0003] In the existing technology, there are deficiencies in advertising budget allocation: traditional advertising budget allocation usually relies on static rules or simple proportional allocation methods, ignoring the dynamic changes in the value differences between advertisements, resulting in the inability to effectively concentrate the budget on high-value advertisements. In addition, fixed thresholds or single indicators are often used as weights, lacking a comprehensive consideration of budget constraints and delivery effects. It is difficult to take into account the dual needs of budget cap restrictions and maximizing advertising value, resulting in low efficiency in advertising budget utilization and limited advertising display effects and conversion benefits. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the present invention provides an advertising effect monitoring and optimization system based on big data to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: In a first aspect, an embodiment of the present invention provides an advertising delivery effect monitoring and optimization system based on big data, comprising the following steps: S1. Extract features based on the advertising log data to obtain an interactive feature set; S2. Predict user behavior based on the interaction feature set to obtain the actual conversion probability; S3. Calculate the advertising conversion value based on the actual conversion probability to obtain the advertising unit value; S4. Sort the advertisements by priority according to the unit value of the advertisements to obtain a sequence of advertisements to be delivered; S5. Calculate the budget allocation based on the sequence of ads to be placed to obtain the advertising budget allocation value. S6. Allocate display opportunities according to the advertising budget allocation value to obtain an display allocation result.

[0006] To further optimize this technical solution, the user behavior prediction in step S2 includes: Based on the obtained interaction feature set, a behavior prediction model is used to predict user behavior by combining click probability and conditional conversion probability to obtain a prediction result of the actual conversion probability.

[0007] To further optimize this technical solution, the behavior prediction model includes:

[0008] in: : actual conversion probability; : click probability; : Conditional conversion probability.

[0009] To further optimize this technical solution, the click probability includes:

[0010] in: : Weight vector of click probability; : The interaction feature vector between users and advertisements; : activation function; By modeling the interaction feature vector, the click probability is calculated.

[0011] To further optimize this technical solution, the conditional conversion probability includes:

[0012] in: : weight vector of conditional conversion probability; Taking click as the prerequisite, the interactive feature vector of the positive sample click behavior is modeled and the conditional conversion probability is calculated.

[0013] To further optimize this technical solution, the advertisement conversion value calculation in step S3 includes: Based on the predicted results of the user's actual conversion probability, by setting a unit conversion value for each ad, based on a single impression, the unit conversion value of the ad is multiplied by the actual conversion probability to calculate the ad conversion value, estimate the expected revenue that the ad display can bring, and obtain the unit value of the ad.

[0014] To further optimize this technical solution, the budget allocation calculation in step S5 includes: Based on the obtained sequence of advertisements to be delivered, a budget allocation model is used to calculate the budget allocation for the advertisements to be delivered, so as to match the budget with the unit value of the advertisements and determine the budget share of each advertisement.

[0015] To further optimize this technical solution, the budget allocation model includes:

[0016]

[0017] in: :advertise The initial budget allocation value of :advertise budget allocation weights; :advertise The actual budget allocation value; : Advertising budget cap; : Total budget.

[0018] To further optimize this technical solution, the budget allocation weights include:

[0019] in: :advertise The unit value of :advertise The unit value of : The total number of ads in the ad sequence to be delivered; By calculating the value ratio of an advertisement in the advertisement sequence, the budget allocation weight of the advertisement is obtained.

[0020] To further optimize this technical solution, the functional modules include: Feature extraction module, probability calculation module, value estimation module, resource allocation module, and display scheduling module.

[0021] In a second aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of a big data-based advertising delivery effect monitoring and optimization system as described in the first aspect of the present invention are implemented.

[0022] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of a big data-based advertising delivery effect monitoring and optimization system as described in the first aspect of the present invention are implemented.

[0023] Compared with the existing technology, the present invention provides an advertising effect monitoring and optimization system based on big data, which has the following beneficial effects: This big data-based advertising effectiveness monitoring and optimization system improves the efficiency of budget utilization, enhances the flexibility and adaptability of budget allocation through a budget allocation model, and sets a budget cap to prevent the single use of resources, reduce risks, and improve the diversity and fairness of advertising. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 This is a flow chart of a system for monitoring and optimizing advertising effectiveness based on big data proposed by the present invention; Figure 2 This is a flow chart of a behavior prediction model for a big data-based advertising effect monitoring and optimization system proposed by the present invention; Figure 3 This is a flow chart of a budget allocation model for a big data-based advertising effectiveness monitoring and optimization system proposed by the present invention; Figure 4 This is a module diagram of a big data-based advertising effect monitoring and optimization system proposed by the present invention. DETAILED DESCRIPTION

[0026] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0027] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0028] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive with other embodiments.

[0029] Example 1: Reference Figures 1 to 3 , which is the first embodiment of the present invention, provides an advertising effect monitoring and optimization system based on big data, including the following steps: S1. Extract features based on the advertising delivery log data to obtain an interactive feature set.

[0030] In this embodiment, the feature extraction includes: In the advertising delivery system, each ad display behavior and conversion result will be recorded in the background in the form of logs. These raw logs contain a large amount of scattered user behavior information and ad attribute information. In order to monitor and optimize the advertising delivery effect, these logs must first be converted into a set of interactive features with clear structure and semantics through feature extraction.

[0031] The purpose of this step is to associate and fuse the multi-source data in the advertising display process to construct a structured feature dataset in a unified format, thereby unifying the semantics and structure of the data and providing a data foundation for subsequent steps.

[0032] The implementation method of this step includes: Log data integration and identification synchronization: Collect exposure logs and conversion logs from the ad delivery system, ensuring that each data entry contains key fields such as user ID, ad ID, display time, click status, and conversion status. Use user unique identifiers and ad unique identifiers to establish associations between logs. Structured design of feature dimensions: Construct three categories of features, including user characteristics (including gender, age, device type, region, historical click behavior, historical conversion behavior, etc.), ad characteristics (including ad category, creative type, delivery time period, exposure platform, etc.), and other characteristics (including current display time, whether it is a holiday, weather, etc.); Feature value encoding and normalization: One-hot encoding is used for discrete features (such as region and ad type). If there are n possible values, each discrete feature data is constructed into an n-dimensional sparse vector with only one position set to 1 and the rest to 0. Numerical features (such as historical click counts) are normalized to a range of 0 to 1, ensuring that all input features meet unified semantic and numerical standards. Definition of label annotation logic: For each exposure record, based on whether it generates conversion behavior within a certain period of time (such as 24 hours) after being displayed, its conversion status is marked, and a set of positive and negative samples is constructed, including those that are converted to positive examples and those that are not converted to negative examples.

[0033] S2. Predict user behavior based on the interaction feature set to obtain the actual conversion probability.

[0034] In this embodiment, the user behavior prediction includes: In step S1, the interactive feature set is obtained to predict the user's click and conversion behavior. Since clicks and conversions are continuous events in the advertising effect chain (users click first, and then conversions may occur), there is an obvious conditional dependency between the two. Therefore, this step constructs a behavior prediction model to predict user behavior by combining click probability and conditional conversion probability, improving prediction accuracy and the consistency of behavior chain modeling, and obtaining a prediction result of the actual conversion probability.

[0035] Furthermore, the behavior prediction model includes:

[0036] in: : Actual conversion probability, which indicates the probability that a certain impression will eventually result in a conversion; : Click probability, which measures the likelihood of an ad display triggering a click and reflects the attractiveness of the ad to users; : Conditional conversion probability, that is, the conditional probability of conversion generated by a clicked ad, reflecting the possibility of actual target behavior such as purchase or registration after clicking.

[0037] Furthermore, the click probability includes:

[0038] in: : The weight vector of click probability, used to calculate the influence of the interaction feature vector on the click probability. It has the same number of dimensions as the interaction feature vector and is initialized with a normal distribution with a mean of 0 and a standard deviation of 0.01. The cross entropy loss function is used (the existing technology is: , is the number of samples, is the click judgment value, 0 means no click, 1 means clicked); : The interaction feature vector between the user and the advertisement, including all encoded and normalized features in the interaction feature set obtained in step S1; : Activation function, Sigmoid activation function, used to make the output range from 0 to 1; By modeling the interaction feature vector, the click probability is calculated.

[0039] Furthermore, the conditional conversion probability includes:

[0040] in: : The weight vector of conditional conversion probability, used to estimate the conversion probability under the premise of clicking. It has the same number of dimensions as the interaction feature vector and is optimized in the same way as the weight vector of click probability, but only uses clicked samples. It is initialized with a normal distribution with a mean of 0 and a standard deviation of 0.01 and uses the cross entropy loss function (existing technology, the formula is , is the number of samples, is the conversion judgment value, 0 means no conversion, 1 means converted) for optimization; Taking click as the prerequisite, the interactive feature vector of the positive sample click behavior is modeled and the conditional conversion probability is calculated.

[0041] The model describes how to use the interactive feature set to estimate the click probability and conditional conversion probability, and finally predict the actual conversion probability.

[0042] In traditional user behavior prediction methods, the behavioral dependency between click probability and conversion probability is often ignored, which may lead to deviations in the joint prediction and make the prediction results less reasonable, interpretable and accurate. However, this model combines the click probability and conversion probability, calculates the conversion probability based on the click probability, which conforms to the behavioral logic. Finally, the actual conversion probability is predicted by combining the click probability and conversion probability, which improves the rationality, interpretability and accuracy of the model results.

[0043] The steps for using this model include: Data acquisition: Get the interaction feature set from step S1 and obtain the interaction feature vector between the user and the advertisement ; Behavior prediction: Based on the obtained interaction feature vector , calculate the click probability and conditional conversion probability ; Conversion probability calculation: Based on the calculated click probability and conditional conversion probability Calculating the actual conversion probability .

[0044] S3. Calculate the advertising conversion value based on the actual conversion probability to obtain the advertising unit value.

[0045] In this embodiment, the advertisement conversion value calculation includes: In step S2, the prediction result of the user's actual conversion probability is obtained. The purpose of this step is to calculate the advertising conversion value based on the predicted actual conversion probability, estimate the expected revenue that can be brought by advertising display, realize the estimation of advertising value, and obtain the advertising unit value, so as to evaluate the potential revenue of advertising to the platform, distinguish between high-value and low-value advertisements, and provide a data basis for the accurate allocation of advertising resources and the optimization of advertising bids.

[0046] The implementation method of this step includes: Behavior probability input processing: the click probability and conditional conversion probability obtained in step S2 are used as input; Target value information introduction: Set a unit conversion value for each ad, which represents the value of each successful conversion of the ad. This value is obtained by using the historical average conversion revenue, a preset CPA value, or the target conversion price set by the advertiser. Calculate the unit impression value: Calculate the expected advertising revenue based on a single impression. Multiply the unit conversion value of the ad by the actual conversion probability to get the unit impression value, which is the expected advertising revenue that each ad impression can bring.

[0047] S4. Sort the advertisements by priority according to the unit value of the advertisements to obtain a sequence of advertisements to be delivered.

[0048] In this embodiment, the advertisement priority sorting includes: In step S3, the expected revenue of the advertisement is evaluated and the unit value of the advertisement is obtained. The purpose of this step is to use the unit value of the advertisement as the sorting basis. Under the premise of meeting the delivery rules (such as time window, user attribute matching, and budget constraints), the currently available advertisements are prioritized to obtain the optimal sequence of advertisements to be delivered, prevent low-value advertisements from occupying exposure opportunities and damaging platform revenue and user experience, achieve the decision-making goal of prioritizing revenue, thereby improving platform resource utilization and advertiser satisfaction, and providing a data basis for subsequent steps.

[0049] The implementation method of this step includes: Input candidate ad sets: Collect the ad sets that currently meet the delivery qualifications and filter them according to multiple conditions (such as target user matching, remaining budget, and display frequency reaching the target); Sorting logic execution: Based on the unit value of each advertisement obtained in step S3, all advertisements are sorted in descending order using a sorting and scheduling engine (such as a real-time priority queue management mechanism based on Apache Flink). Higher values ​​give higher priorities. Generate ad sequences: After sorting is complete, output a set of ad sequences prioritized by expected revenue for use in subsequent steps.

[0050] S5. Calculate the budget allocation based on the sequence of advertisements to be delivered to obtain an advertisement budget allocation value.

[0051] In this embodiment, the budget allocation calculation includes: In step S4, the sequence of advertisements to be delivered is obtained. The sequence is prioritized according to the contribution of each advertisement to the expected revenue of the platform during the delivery cycle. In order to maximize revenue, the purpose of this step is to use the budget allocation model to calculate the budget allocation for the advertisements to be delivered within a given budget and delivery cycle, so as to match the budget with the unit value of the advertisement and determine the budget share of each advertisement.

[0052] Furthermore, the budget allocation model includes:

[0053]

[0054] in: :advertise The initial budget allocation value of :advertise The budget allocation weight of represents the value of an ad in the current ad sequence and determines the budget share it should receive. :advertise The actual budget allocation value; Ad budget cap: This indicates the maximum budget available for each ad. It is set based on platform restrictions to prevent single-use of resources, excessive consumption of ad budgets, or exceeding the budget affordability. : Total budget, which indicates the total budget available within a delivery cycle, is usually set by the platform based on the funding allocation of the delivery cycle.

[0055] Furthermore, the budget allocation weights include:

[0056] in: :advertise The unit value of advertising The expected advertising value of a unit display opportunity is obtained in step S3; :advertise The unit value of advertising The expected advertising value of a unit display opportunity is obtained in step S3; : The total number of ads in the ad sequence to be delivered; By calculating the value ratio of an advertisement in the advertisement sequence, the budget allocation weight of the advertisement is obtained.

[0057] This model describes how to allocate advertising budget according to advertising priority to maximize revenue.

[0058] Traditional budget allocation methods usually allocate budgets evenly or linearly according to the budget cap statically set by advertisers, without considering conversion quality or unit benefits. This makes it difficult to dynamically respond to fluctuations in advertising value, resulting in resource waste and low budget utilization. This model allocates budgets based on advertising priority, improving resource utilization, enhancing the flexibility and adaptability of budget allocation, and setting a budget cap to prevent the single use of resources, reduce risks, and improve advertising diversity and fairness.

[0059] The steps for using the above model include: Data acquisition: Get the sequence of ads to be delivered from step S4, and get the unit value of all ads in the sequence of ads to be delivered from step S3 ; Allocation ratio calculation: Calculate the budget allocation weight of each advertisement based on the unit value of the advertisement , get the initial proportion of each ad in the total budget; Budget allocation: Based on the total advertising budget of the platform , combined with the budget allocation weight of each ad , calculate the initial budget allocation for each ad , to determine whether each ad exceeds the budget limit , if it exceeds, take the minimum value to get the actual budget allocation value If the total budget is not fully allocated because there are ads that exceed the budget cap, the allocation ratio will be recalculated and redistributed among the ads that do not reach the cap.

[0060] S6. Allocate display opportunities according to the advertising budget allocation value to obtain an display allocation result.

[0061] In this embodiment, the display opportunity allocation includes: In the advertising delivery system, the platform controls the display opportunities (i.e., exposure positions) generated each time a user visits or refreshes the platform. These display opportunities are limited in number and need to be allocated based on estimated value. Therefore, it is necessary to allocate display opportunities based on the advertising budget ratio obtained in step S5, and reasonably allocate display positions to the optimal advertising combination, thereby increasing the commercial value of each exposure, maximizing platform revenue and budget utilization efficiency, and preventing certain ads from losing display opportunities, thereby ensuring fairness in advertising display.

[0062] The implementation method of this step includes: Budget share calculation: Calculate the share of each advertisement in the total budget based on the advertising budget allocation value obtained in step S5; Display pool creation: Build a display pool based on the total number of display slots available in the current time period (for example, there are 1,000 display opportunities in 1 second); Display opportunity allocation: Display opportunities are allocated to each ad based on their budget share. For example, if Ad A's budget share is 30%, it will theoretically receive 30% of the display positions. Queuing candidate strategy processing: If an ad is delivered frequently but has a high user repetition rate, its priority will be temporarily lowered to prevent a degradation in user experience.

[0063] Example 2: Reference Figure 4 , which is the second embodiment of the present invention, provides an advertising effect monitoring and optimization system based on big data, including the following functional modules: Feature extraction module: used to extract the interaction features between users and ads based on the ad placement log data to obtain the interaction feature set; Probability calculation module: Based on the interaction feature set between users and ads, it predicts user behavior and calculates the actual conversion probability of users to ads; Value estimation module: This module conducts a comprehensive evaluation based on the historical conversion revenue and expected revenue of an ad, and maps the predicted actual conversion rate into the unit value of the ad. Resource allocation module: allocates budget based on the unit value of the ad and the budget cap, ensuring that high-value ads are given priority and that each ad receives budget allocation; Display Scheduling Module: Calculates the budget share of each ad based on the allocated budget, allocates display opportunities, and controls display frequency, enhancing user experience and advertiser satisfaction while ensuring overall revenue.

[0064] Example 3: This embodiment also provides a computer device, which is suitable for an advertising delivery effect monitoring and optimization system based on big data, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement an advertising delivery effect monitoring and optimization system based on big data as proposed in the above embodiment.

[0065] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, an advertising delivery effect monitoring and optimization system based on big data as proposed in the above embodiment is implemented.

[0066] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0067] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0068] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0069] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0070] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field-programmable gate array (FPGA), etc.

[0071] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A big data-based advertising effect monitoring and optimization system, characterized in that: The following steps are involved: S1. Extract features based on the advertising log data to obtain an interactive feature set; S2. Predict user behavior based on the interaction feature set to obtain the actual conversion probability; S3. Calculate the advertising conversion value based on the actual conversion probability to obtain the advertising unit value; S4. Sort the advertisements by priority according to the unit value of the advertisements to obtain a sequence of advertisements to be delivered; S5. Calculate the budget allocation based on the sequence of ads to be placed to obtain the advertising budget allocation value. S6. Allocate display opportunities according to the advertising budget allocation value to obtain an display allocation result.

2. The system for monitoring and optimizing advertising effectiveness based on big data according to claim 1, characterized in that: The user behavior prediction in step S2 includes: Based on the obtained interaction feature set, a behavior prediction model is used to predict user behavior by combining click probability and conditional conversion probability to obtain a prediction result of the actual conversion probability.

3. The big data-based advertising effect monitoring and optimization system according to claim 2, characterized in that: The behavior prediction model includes: ， in: : actual conversion probability; : click probability; : Conditional conversion probability.

4. The system for monitoring and optimizing advertising effectiveness based on big data according to claim 3, characterized in that: The click probability includes: ， in: : Weight vector of click probability; : The interaction feature vector between users and advertisements; : activation function; By modeling the interaction feature vector, the click probability is calculated.

5. The system for monitoring and optimizing advertising effects based on big data according to claim 3, characterized in that: The conditional conversion probability includes: ， in: : weight vector of conditional conversion probability; Taking click as the prerequisite, the interactive feature vector of the positive sample click behavior is modeled and the conditional conversion probability is calculated.

6. The system for monitoring and optimizing advertising effectiveness based on big data according to claim 1, characterized in that: The advertisement conversion value calculation in step S3 includes: Based on the predicted results of the user's actual conversion probability, by setting a unit conversion value for each ad, based on a single impression, the unit conversion value of the ad is multiplied by the actual conversion probability to calculate the ad conversion value, estimate the expected revenue that the ad display can bring, and obtain the unit value of the ad.

7. The system for monitoring and optimizing advertising effectiveness based on big data according to claim 1, characterized in that: The budget allocation calculation in step S5 includes: Based on the obtained sequence of advertisements to be delivered, a budget allocation model is used to calculate the budget allocation for the advertisements to be delivered, so as to match the budget with the unit value of the advertisements and determine the budget share of each advertisement.

8. The system for monitoring and optimizing advertising effects based on big data according to claim 7, characterized in that: The budget allocation model includes: ， ， in: :advertise The initial budget allocation value of :advertise budget allocation weights; :advertise The actual budget allocation value; : Advertising budget cap; : Total budget.

9. The big data-based advertising effect monitoring and optimization system according to claim 8, characterized in that: The budget allocation weights include: ， in: :advertise The unit value of :advertise The unit value of : The total number of ads in the ad sequence to be delivered; By calculating the value ratio of an advertisement in the advertisement sequence, the budget allocation weight of the advertisement is obtained.

10. The system for monitoring and optimizing advertising effects based on big data according to claim 1, characterized in that: The functional modules of the system include: Feature extraction module, probability calculation module, value estimation module, resource allocation module, and display scheduling module.

Images