Advertisement delivery method based on user portrait

Abstract

The application relates to the technical field of targeted advertising, in particular to an advertisement putting method based on user portraits, which comprises the following steps: obtaining the equipment use time of a target user, recording each software use type and corresponding use frequency and corresponding advertisement response information, dividing the equipment use time into a plurality of advertisement putting characteristic time periods, and determining the advertisement putting type of each advertisement putting characteristic time period; according to the advertisement feedback attitude of the user in each characteristic time period, determining the advertisement effective tendency category of the current user in each advertisement putting characteristic time period, so as to determine the basic advertisement putting mode; according to the time length percentage in each time period, determining the advertisement acceptance fluctuation parameter of the user in the current time period, so as to determine whether to adjust the advertisement putting content of the current characteristic time period, and according to the advertisement watching time length percentage of each use type software, determining the adjustment direction; the application can accurately reach different time period users with advertisement resources, and improves the overall advertisement putting effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of targeted advertising, and in particular to an advertisement delivery method based on user portraits. BACKGROUND

[0002] Advertisement delivery is supported by big data and artificial intelligence, and user basic information and behavior data are obtained through multi-channel data collection technology (covering social media interaction, browsing track, consumption record, etc.), and cross-scene behavior is connected by combining device fingerprint, browser tracking and other tools; dynamic user portraits are constructed by machine learning algorithms (such as clustering analysis and feature engineering), and features such as interest tags, consumption ability and potential demand are extracted from massive data, and portrait dimensions are iterated in real time relying on AI models; in the delivery link, precise coupling of advertisement content and user features is realized by using keyword matching and interest targeting algorithms, and real-time bidding (RTB) systems are used to dynamically adjust delivery strategies, and data utilization and privacy protection are balanced within the compliance framework through differential privacy and federated learning technologies, and finally a full-process technical closed loop from data collection, portrait construction to precise reach is formed.

[0003] Chinese Patent Publication No. CN116805255B discloses an advertisement automatic optimization delivery system based on user portrait analysis, which includes an intelligent push module, a user portrait module, a delivery optimization module and an effect analysis module. The user portrait module is used to label users according to shopping information to obtain a user portrait model of the user. The intelligent push module is used to automatically deliver advertisements to users based on the user portrait model. The effect analysis module is used to analyze the traffic effect of the goods after the delivery of the advertisements, and generate a poor benefit signal, a slow benefit signal or a normal benefit signal. The delivery optimization module is used to receive delivery optimization instructions to optimize the delivery of the corresponding advertisements of the goods. It can be seen that the advertisement automatic optimization delivery system based on user portrait analysis has the following problems: the actual needs of the delivered users at different times are not considered, and the delivery is only based on the content, resulting in inaccurate delivery at different time periods. SUMMARY

[0004] Therefore, the present application provides an advertisement delivery method based on user portraits to overcome the problem in the prior art that the actual needs of the delivered users at different times are not considered, and the delivery is only based on the content, resulting in inaccurate delivery at different time periods.

[0005] To achieve the above-mentioned purpose, as a preferred technical solution of the advertisement delivery method based on user portraits, it comprises:

[0006] Obtaining device usage time of a target user, software usage types of each software and usage frequency of each software, and advertisement response information in the process of using each software, wherein the advertisement response information includes a watching time percentage and an advertisement feedback attitude;

[0007] According to the usage type of each software combined with the corresponding usage frequency, the device usage time is divided into several delivery feature time periods, and the advertisement delivery type of each delivery feature time period is determined based on the software usage type combined with the usage frequency;

[0008] According to the advertisement feedback attitude of the user in each feature time period, the advertisement effective tendency category of the current user in each delivery feature time period is determined to determine the basic advertisement delivery method, including reverse advertisement delivery or targeted delivery according to the advertisement delivery type of the feature time period;

[0009] According to the time percentage monitored in real time in each delivery feature time period, the advertisement acceptance fluctuation parameter of the user in the current time period is determined to determine whether to adjust the advertisement delivery content of the current feature time period, and the adjustment direction is determined according to the advertisement watching time percentage of each usage type software.

[0010] As a preferred technical solution of the advertisement delivery method based on user portrait, the device usage time is divided into several feature time periods according to the usage type of each software and the corresponding usage frequency, including:

[0011] According to the total usage frequency of all software, a single delivery period is divided into several delivery feature time periods;

[0012] According to the usage type of the software with the highest usage frequency in each feature time period, the advertisement delivery type of the current feature time period is determined;

[0013] Among them, the time length of each delivery feature time period is different.

[0014] As a preferred technical solution of the advertisement delivery method based on user portrait, the delivery feature time periods are divided into three types according to the total software usage frequency, including non-active delivery time period, active delivery time period and general delivery time period;

[0015] If the total usage frequency of all software is less than the non-active frequency threshold, the current delivery feature time period is a non-active delivery time period;

[0016] If the total usage frequency of all software is greater than or equal to the active frequency threshold, the current delivery feature time period is an active delivery time period;

[0017] If the total usage frequency of the total software is greater than or equal to the inactive frequency threshold and less than the active frequency threshold, the current delivery feature time period is a general delivery time period.

[0018] As a preferred technical solution of the advertisement delivery method based on the user portrait, the advertisement effective tendency category of the current user in each delivery feature time period is determined according to the advertisement feedback attitude of the target user in each feature time period, comprising:

[0019] Determine the advertisement feedback attitude of the target user each time, the advertisement feedback attitude includes positive feedback and negative feedback;

[0020] According to the proportion of the number of negative feedbacks, the advertisement effective tendency category of the current time period is determined;

[0021] If the number of negative feedbacks is greater than the preset negative feedback threshold, the advertisement effective tendency category of the current time period is an advertisement delivery invalid category;

[0022] If the number of negative feedbacks is less than or equal to the preset negative feedback threshold, the advertisement effective tendency category of the current time period is an advertisement delivery effective category.

[0023] As a preferred technical solution of the advertisement delivery method based on the user portrait, the basic advertisement delivery mode is determined according to the advertisement effective tendency category of the current feature time period, comprising:

[0024] If the advertisement effective tendency category is an advertisement delivery effective category, the basic advertisement delivery mode is to perform targeted delivery according to the advertisement delivery type of the feature time period;

[0025] If the advertisement effective tendency category is an advertisement delivery invalid category, the basic advertisement delivery mode is to perform reverse advertisement delivery.

[0026] As a preferred technical solution of the advertisement delivery method based on the user portrait, performing reverse advertisement delivery comprises:

[0027] In response to the user in the corresponding active delivery time period, an advertisement different from the user information record is delivered, and its advertisement feedback attitude is recorded, and the delivery is replaced in the direction from large to small according to the difference.

[0028] As a preferred technical solution of the advertisement delivery method based on the user portrait, the advertisement acceptance fluctuation parameter in the current delivery feature time period is determined, comprising:

[0029] Determine the viewing time percentage of all advertisements in the current delivery feature time period and calculate the average value;

[0030] Calculate the average deviation of the viewing time percentage of all advertisements;

[0031] The advertisement acceptance fluctuation parameter is determined according to a ratio of the average deviation to the percentage of the watching time and the average value.

[0032] As the preferred technical scheme of the advertisement putting method based on the user portrait, whether to adjust the advertisement putting content of the current characteristic time period is determined according to the advertisement acceptance fluctuation parameter, and the method comprises the following steps of:

[0033] If the advertisement acceptance fluctuation parameter is less than or equal to a standard acceptance fluctuation parameter, it is determined that the advertisement putting content of the current characteristic time period is not adjusted.

[0034] If the advertisement acceptance fluctuation parameter is greater than the standard acceptance fluctuation parameter, it is determined that the advertisement putting content of the current characteristic time period is adjusted.

[0035] As the preferred technical scheme of the advertisement putting method based on the user portrait, the adjustment direction is determined according to the percentage of the advertisement watching time of each software using type, and the method comprises the following steps of:

[0036] The software using type with the maximum average value of the percentage of the watching time is determined as the optimal advertisement putting type.

[0037] The put advertisement of the software of the remaining type is associated with the optimal advertisement putting type, and is put into the corresponding software.

[0038] The present application has the following beneficial effects:

[0039] The application can effectively distinguish the effective and ineffective categories of advertisement delivery by analyzing the advertisement feedback attitude of the target user in detail, accurately positioning the effective tendency categories of the advertisement in each characteristic time period, determining the basic advertisement delivery mode based on the above, and setting a reasonable negative feedback threshold, so as to realize accurate advertisement resource delivery and enhance the efficiency of advertisement delivery. For the active delivery time period, the advertisement with a large difference from the information obtained by the user is delivered, the information cocoon is avoided, the freshness and attractiveness of the advertisement are improved, and the click and conversion are increased. In the non-active and general time period, the reverse advertisement delivery strategy is adopted, the advertisement content recommendation mode is changed, the advertisement attention is improved again, the delivery efficiency is optimized, the advertisement resource accurately reaches the users in different time periods, and the overall advertisement delivery effect is improved. The average value and average deviation of the watching time percentage are calculated, and the ratio of the two is further calculated to determine the advertisement acceptance fluctuation parameter, which can accurately reflect the dispersion degree of the user watching behavior, effectively avoid the misjudgment caused by the difference in user watching habits, and comprehensively improve the advertisement delivery effect and user acceptance. By quantitatively analyzing the advertisement watching time percentage of different software usage types, the software type that the user is most interested in is accurately identified as the optimal advertisement delivery type, and the advertisements of other types of software are associated and delivered, which can accurately position the user interest point, improve the relevance of advertisement delivery, and thus improve the click rate and conversion rate of the advertisement, optimize the allocation of advertisement resources, and improve the overall delivery effect.

[0040] Especially, in the application, by analyzing the advertisement feedback attitude of the target user in detail, accurately positioning the effective tendency categories of the advertisement in each characteristic time period, and determining the basic advertisement delivery mode based on the above, in combination with the negative feedback behavior of the user, the effective and ineffective categories of advertisement delivery can be effectively distinguished by setting a reasonable negative feedback threshold. When the feedback is positive, targeted delivery is implemented to improve the advertisement effect; when the negative feedback exceeds the threshold, the reverse delivery strategy is started, new touch content is generated by means of adversarial training, and traditional ineffective traffic is avoided, so as to optimize the delivery efficiency, realize accurate delivery of advertisement resources, and enhance the efficiency of advertisement delivery.

[0041] Especially, in the application, for the active delivery time period, the advertisement with a large difference from the information obtained by the user is delivered, the information cocoon is avoided, the freshness and attractiveness of the advertisement are improved, and the click and conversion are increased. The advertisement of the same type but with a changed recommendation mode, such as an opposite tone, picture and copy, is delivered according to the difference from large to small, the advertisement attention is improved again, the traditional ineffective traffic is avoided, the delivery efficiency is optimized, the advertisement resource accurately reaches the users in different time periods, and the overall advertisement delivery effect is improved.

[0042] Especially, in the application, the fluctuation parameter of advertisement acceptance is determined by calculating the average value and average deviation of the percentage of watching time, and further calculating the ratio of the two, which can accurately reflect the dispersion degree of user watching behavior, effectively avoid misjudgment caused by differences in user watching habits, and the parameter does not simply depend on fixed watching time, but considers the dispersion degree of the percentage of watching time, can more comprehensively reflect the average satisfaction of users to the advertisement, and improve the advertisement delivery effect.

[0043] Especially, in the application, by quantitatively analyzing the advertisement watching time percentage of different software usage types, the software type that the user is most interested in is accurately identified as the optimal advertisement delivery type, and the advertisements of other types of software are associated and delivered, which can accurately locate the user's interest point, improve the relevance of advertisement delivery, and thus improve the click rate and conversion rate of the advertisement; in addition, it can also optimize the allocation of advertisement resources, tilt more resources to the software type that the user pays more attention to, avoid waste in inefficient types, and improve the overall delivery effect. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 The flowchart of the advertisement delivery method based on user portrait of the embodiment of the application;

[0045] Figure 2 The logic diagram for determining the effective tendency category of the advertisement of the embodiment of the application;

[0046] Figure 3 The logic diagram for determining whether to adjust the advertisement delivery content of the current feature time period of the embodiment of the application. DETAILED DESCRIPTION

[0047] In order to make the purpose and advantages of the application more clear and obvious, the application will be further described below in combination with embodiments; it should be understood that the specific embodiments described herein are only used to explain the application, and not to limit the application.

[0048] The preferred embodiments of the application will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the application, and are not intended to limit the protection scope of the application.

[0049] It should be noted that in the description of the application, the terms "up", "down", "left", "right", "in", "out" and the like indicate the direction or positional relationship terms based on the direction or positional relationship shown in the drawings, which are only for the convenience of description, and do not indicate or imply that the device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the application.

[0050] Please refer to Figure 1As shown, it is a flow chart of the advertisement putting method based on user portrait according to the embodiment of the present application; the present application provides an advertisement putting method based on user portrait, which comprises:

[0051] In step S1, the device usage time of the target user, the software usage type of each software, the usage frequency of each software, and the advertisement response information in the process of using each software are obtained, wherein the advertisement response information comprises the watching time percentage and the advertisement feedback attitude;

[0052] In step S2, the device usage time is divided into several putting feature time periods according to the usage type of each software combined with the corresponding usage frequency, and the advertisement putting type of each putting feature time period is determined based on the software usage type combined with the usage frequency;

[0053] In step S3, the advertisement effective tendency category of the current user in each putting feature time period is determined according to the advertisement feedback attitude of the user in each feature time period, so as to determine the basic advertisement putting mode, which comprises: carrying out the advertisement putting in reverse, or carrying out the targeted putting according to the advertisement putting type of the feature time period;

[0054] In step S4, the advertisement acceptance fluctuation parameter of the user in the current time period is determined according to the watching time percentage monitored in each putting feature time period, so as to determine whether to adjust the advertisement putting content of the current feature time period, and the adjustment direction is determined according to the advertisement watching time percentage of each usage type software.

[0055] In the implementation, the software usage type comprises social entertainment, tool office, life service and education culture.

[0056] It should be noted that the information of the target user (including but not limited to the device information of the account, the personal information of the user, etc.), the related data, etc. involved in the present embodiment are all the information authorized by the account or authorized by each party.

[0057] The application can effectively distinguish the effective and ineffective categories of advertisement delivery by analyzing the advertisement feedback attitude of target users, accurately positioning the effective tendency categories of each feature time period, determining the basic advertisement delivery mode based on the above, and setting a reasonable negative feedback threshold, so as to realize accurate advertisement resource delivery and enhance the efficiency of advertisement delivery; for active delivery time periods, advertisements with large differences between delivery and user information acquisition are delivered to avoid information cocooning, improve the freshness and attractiveness of advertisements, and increase clicks and conversions; in non-active and general time periods, a reverse advertisement delivery strategy is adopted to change the advertisement content recommendation mode, re-increase the attention of advertisements, optimize the delivery efficiency, make the advertisement resources accurately reach different time period users, and improve the overall advertisement delivery effect; by calculating the average value and average deviation of the percentage of watching time and further calculating the ratio to determine the advertisement acceptance fluctuation parameter, the discrete degree of user watching behavior can be accurately reflected, the misjudgment caused by the difference in user watching habits can be effectively avoided, and the advertisement delivery effect and user acceptance can be comprehensively improved; by quantitatively analyzing the percentage of advertisement watching time of different software usage types, the most interested software type of the user is accurately identified as the optimal advertisement delivery type, and the advertisements of other types of software are associated and delivered, which can accurately position the user's interest point, improve the relevance of advertisement delivery, and thus improve the click rate and conversion rate of advertisements, optimize the allocation of advertisement resources, and improve the overall delivery effect.

[0058] Specifically, in the step S2, the device usage time is divided into several feature time periods according to the usage type and corresponding usage frequency of each software, including:

[0059] According to the total usage frequency of all software, a single delivery period is divided into several delivery feature time periods;

[0060] According to the usage type of the software with the highest usage frequency in each feature time period, the advertisement delivery type of the current feature time period is determined;

[0061] Among them, the time length of each delivery feature time period is different.

[0062] It can be understood that for a single user, the use type and frequency of each software in the device will change over time, for example, during working hours, the use frequency of social entertainment type software is relatively low, the use frequency of tool office type software is relatively high, and the total software use frequency is also relatively low, and conversely, for non-working time such as leisure time at night, the user uses social entertainment type \ life service type and education and culture type software The frequency will increase significantly, and at this time the total software use frequency will also increase. During the non-active time, such as the deep sleep stage, the user will hardly use any software actively, and the device is usually in standby or locked state, and the software use frequency and use time are reduced to a very low level. Relatively, in the stage of difficulty falling asleep or about to fall asleep, some users may use some sleep aid software such as light music player and white noise application for a short time, so the use frequency of this part of the feature time period is not 0.

[0063] Specifically, the feature time period includes three types of total software use frequency, including non-active feature time period, active feature time period, and general feature time period.

[0064] If the total use frequency of all software is less than the non-active frequency threshold, the current feature time period is a non-active feature time period.

[0065] If the total use frequency of all software is greater than or equal to the active frequency threshold, the current feature time period is an active feature time period.

[0066] If the total use frequency of all software is greater than or equal to the non-active frequency threshold and less than the active frequency threshold, the current feature time period is a general feature time period.

[0067] In implementation, the non-active frequency threshold is initially set to less than 5 times per hour, and the active frequency threshold is initially set to greater than or equal to 20 times per hour. Each threshold can be adaptively adjusted according to the actual use frequency and life time of the target user, and each target user can be divided according to the actual use condition, without specific limitation.

[0068] It can be understood that generally, during the non-active time period such as deep sleep or busy work, the software use frequency will be very low. For example, during the late night to early morning period, most people are in a sleep state and hardly use software. A lower threshold can effectively identify these non-active time periods and avoid unnecessary types of advertisement delivery during these periods, thereby reducing the disturbance to the user and improving the effectiveness and pertinence of advertisement delivery.

[0069] During active time periods such as user leisure time, entertainment time, or work break time on weekdays, the user has more free time and higher willingness to use various types of software. For example, from 7 pm to 10 pm, people are mostly in a relaxed state and will frequently use various types of software such as social, entertainment, shopping, etc. During the lunch break on weekdays, some users will also use this time to browse news, play music, etc. A higher active frequency threshold can accurately capture the peak period of software use by these users, and placing advertisements during this period can increase the exposure rate and click rate of the advertisements and effectively reach the target users.

[0070] The use frequency corresponding to the general placement time period is between non-active and active, and the software use frequency of the user in this time period is relatively moderate, and there is a certain software use demand, but it does not reach a high active degree. For example, during the morning and afternoon work time on weekdays, the user will occasionally use some work-related tool software, but the use frequency is relatively low; or during the daytime on weekends, the user will intermittently use software when performing some housework or other leisure activities. Placing advertisements for this time period can select some advertisement contents related to the user behavior characteristics of this time period according to the target of the advertiser to achieve a better placement effect.

[0071] Please refer to Figure 2 The logic diagram for determining the effective tendency category of the advertisement by the embodiment of the present application is shown in FIG. 1, and in step S3, the effective tendency category of the advertisement of the current user in each placement characteristic time period is determined according to the advertisement feedback attitude of the target user in each characteristic time period, which includes:

[0072] The advertisement feedback attitude of the target user is determined, and the advertisement feedback attitude includes positive feedback and negative feedback;

[0073] The effective tendency category of the advertisement of the current time period is determined according to the proportion of the number of negative feedbacks;

[0074] If the number of negative feedbacks is greater than a preset negative feedback threshold, the effective tendency category of the advertisement of the current time period is an invalid category of advertisement placement;

[0075] If the number of negative feedbacks is less than or equal to the preset negative feedback threshold, the effective tendency category of the advertisement of the current time period is an effective category of advertisement placement.

[0076] In implementation, negative feedback is that the advertisement stay time is less than 2 s, or the option of not wanting to see again is selected, and the negative feedback does not conform to the positive feedback of the advertisement.

[0077] The preset negative feedback threshold is determined according to the industry report in the field of e-commerce advertisements, and the negative feedback threshold is usually set to 8% to 15%. Preferably, 11.5% is used as the preset negative feedback threshold.

[0078] Specifically, the basic advertisement launching mode is determined according to the advertisement effective tendency category of the current feature time period, and the basic advertisement launching mode comprises:

[0079] If the advertisement effective tendency category is an advertisement launching effective category, the basic advertisement launching mode is targeted launching according to the advertisement launching type of the feature time period.

[0080] If the advertisement effective tendency category is an advertisement launching ineffective category, the basic advertisement launching mode is reverse advertisement launching.

[0081] It can be understood that, based on the negative feedback data (such as shielding frequency and exposure termination speed) of the user advertisement avoidance behavior, the counter-training mechanism is used to generate counterintuitive touch content. The traditional ineffective flow is avoided, and the overall launching efficiency coefficient is optimized.

[0082] In the present application, by analyzing the advertisement feedback attitude of the target user in detail, the advertisement effective tendency category of each feature time period is accurately positioned, and the basic advertisement launching mode is determined based on the same. By setting a reasonable negative feedback threshold and combining the negative feedback behavior of the user, the advertisement launching effective and ineffective categories can be effectively distinguished. When the feedback is positive, targeted launching is implemented to improve the advertisement effect; when the negative feedback exceeds the threshold, the reverse launching strategy is started, new touch content is generated by means of counter-training, traditional ineffective flow is avoided, and the launching efficiency is optimized, so as to realize the precise launching of advertisement resources and enhance the efficiency of advertisement launching.

[0083] Specifically, the reverse advertisement launching comprises:

[0084] In response to the user in the corresponding active launching time period, an advertisement different from the information record obtained by the user is launched, and the advertisement feedback attitude is recorded, and the launching is replaced in the direction from large to small difference.

[0085] In implementation, for the reverse advertisement launching of the non-active launching time period and the general launching time period, the launched advertisement type is the same as the advertisement launching type of the current feature time period of the user, but the advertisement content recommendation mode is different.

[0086] The different advertisement content recommendation modes include that the tone, picture, and script of the advertisement are opposite to the common content of the type of advertisement.

[0087] In the present application, for the active delivery time period, the advertisement with large difference in user information acquisition is delivered, which can avoid information cocoon; the delivery is replaced according to the difference from large to small, the freshness and attraction of the advertisement are improved, and the click and conversion are increased. In the non-active and general time period, the same advertisement type is maintained, but the recommendation method is changed, such as using opposite tone, picture and copy, which can improve the attention of the advertisement again; at the same time, this method avoids traditional invalid traffic, optimizes the delivery efficiency, makes the advertisement resource accurately reach the users in different time periods, and improves the overall advertisement delivery effect.

[0088] Specifically, in the step S4, the advertisement acceptance fluctuation parameter in the current delivery characteristic time period is determined, including:

[0089] The viewing time percentage of all advertisements in the current delivery characteristic time period is determined and the average value is calculated;

[0090] The average deviation of the viewing time percentage of all advertisements is calculated;

[0091] The ratio of the average deviation to the viewing time percentage and the average value is calculated to determine the advertisement acceptance fluctuation parameter.

[0092] The advertisement acceptance fluctuation parameter is used to represent the discrete degree of the advertisement viewing time distribution.

[0093] Please refer to Figure 3 The logic diagram for determining whether to adjust the advertisement delivery content of the current characteristic time period is shown in the figure, according to the advertisement acceptance fluctuation parameter, whether to adjust the advertisement delivery content of the current characteristic time period is determined, including:

[0094] If the advertisement acceptance fluctuation parameter is less than or equal to the standard acceptance fluctuation parameter, it is determined that the advertisement delivery content of the current characteristic time period is not adjusted;

[0095] If the advertisement acceptance fluctuation parameter is greater than the standard acceptance fluctuation parameter, it is determined that the advertisement delivery content of the current characteristic time period is adjusted.

[0096] In the implementation, the standard acceptance fluctuation parameter is determined according to the interval formed by the difference between the average value of the advertisement acceptance fluctuation parameter in the delivery characteristic time period of the same type and the two times of the standard deviation in the historical record and the two times of the standard deviation. In statistics, about 95% of the data in the normal distribution is located in the interval of the average value ± 2 times of the standard deviation. With this interval as a reference, the interference of extreme abnormal values can be effectively filtered out, so that the standard is more stable and reliable.

[0097] It can be understood that the advertisement acceptance fluctuation parameter is used to determine whether to adjust the delivery content, which is based on its accurate reflection of the dispersion degree of user viewing behavior, and the measurement method can effectively capture the average satisfaction of users to the advertisement, and avoid misjudgment due to differences in user viewing habits (such as accelerated playback, only half watching, and jumping watching, etc.). Different user viewing habits will lead to different viewing time of the same advertisement in different users, and the advertisement acceptance fluctuation parameter calculates the dispersion degree of the viewing time percentage, rather than simply relying on fixed time, which can more comprehensively reflect the acceptance stability and overall satisfaction of users to the advertisement content. Such a measurement standard enables advertisers to adjust the advertising strategy in a timely and accurate manner according to actual user feedback, so as to improve the delivery effect and user acceptance of the advertisement.

[0098] In the present application, the advertisement acceptance fluctuation parameter is determined by calculating the average value and average deviation of the viewing time percentage, and further calculating the ratio thereof, which can accurately reflect the dispersion degree of user viewing behavior and effectively avoid misjudgment due to differences in user viewing habits. This parameter does not simply rely on fixed viewing time, but considers the dispersion degree of the viewing time percentage, which can more comprehensively reflect the average satisfaction of users to the advertisement and improve the advertisement delivery effect.

[0099] Specifically, the adjustment direction is determined according to the advertisement viewing time percentage of each software usage type, including:

[0100] The software usage type with the maximum average value of the viewing time percentage is determined as the optimal advertisement delivery type;

[0101] The delivery advertisement of the remaining types of software is associated with the optimal advertisement delivery type and delivered to the corresponding software.

[0102] In the present application, the advertisement viewing time percentage of different software usage types is quantitatively analyzed to accurately identify the software type that users are most interested in as the optimal advertisement delivery type, and the advertisements of other types of software are associated and delivered, which can accurately locate the user interest point, improve the relevance of advertisement delivery, and thus improve the click-through rate and conversion rate of the advertisement. In addition, it can also optimize the allocation of advertisement resources, tilt more resources to the software types that users pay more attention to, avoid waste in inefficient types, and improve the overall delivery effect.

[0103] If the user portrait-based advertisement delivery method of the present application is realized in the form of a software function 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 solutions of the present application, essentially or in other words, the part that contributes to the prior art, or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0104] The flowcharts and block diagrams in the drawings illustrate the possible implementation architectures, functions and operations of the apparatuses, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowcharts or block diagrams can represent a module, a program segment or a part of code that includes one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur in different orders than those noted in the drawings. For example, two blocks that are represented in succession can actually be executed in parallel, and sometimes in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and the combination of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based device that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0105] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after these changes or replacements will all fall within the protection scope of the present application.

[0106] The above description is only for the preferred embodiments of the present application and is not intended to limit the present application; for those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for delivering advertisements based on user portraits, characterized in that: The method comprises the following steps: Obtaining the device usage time of the target user, the software usage type of each software, the usage frequency of each software, and the advertisement response information in the process of using each software, wherein the advertisement response information comprises the viewing time percentage and the advertisement feedback attitude; Dividing the device usage time into several advertisement delivery feature time periods according to the usage type of each software combined with the corresponding usage frequency, and determining the advertisement delivery type of each advertisement delivery feature time period based on the software usage type combined with the usage frequency; Determining the advertisement effective tendency category of the current user in each advertisement delivery feature time period according to the advertisement feedback attitude of the user in each feature time period, so as to determine the basic advertisement delivery mode, including: performing reverse advertisement delivery, or performing targeted delivery according to the advertisement delivery type of the feature time period; Determining the advertisement acceptance fluctuation parameter of the user in the current time period according to the real-time monitored time percentage in each advertisement delivery feature time period, so as to determine whether to adjust the advertisement delivery content of the current feature time period, and determining the adjustment direction according to the advertisement viewing time percentage of each usage type software; Determining the advertisement acceptance fluctuation parameter in the current advertisement delivery feature time period comprises: Determining the viewing time percentage of all advertisements in the current advertisement delivery feature time period and calculating the average value thereof; Calculating the average deviation of the viewing time percentage of all advertisements; Determining the advertisement acceptance fluctuation parameter according to the ratio of the average deviation to the viewing time percentage and the average value thereof; Determining whether to adjust the advertisement delivery content of the current feature time period according to the advertisement acceptance fluctuation parameter comprises: If the advertisement acceptance fluctuation parameter is less than or equal to the standard acceptance fluctuation parameter, it is determined that the advertisement delivery content of the current feature time period is not adjusted; If the advertisement acceptance fluctuation parameter is greater than the standard acceptance fluctuation parameter, it is determined that the advertisement delivery content of the current feature time period is adjusted; Determining the adjustment direction according to the advertisement viewing time percentage of each usage type software comprises: Determining the software usage type with the maximum average value of the viewing time percentage as the optimal advertisement delivery type; Associating the delivery advertisements of the remaining types of software with the optimal advertisement delivery type and delivering them to the corresponding software. 2.The user portrait-based advertisement distribution method according to claim 1, wherein, Dividing the device usage time into several feature time periods according to the usage type of each software and the corresponding usage frequency comprises: Dividing a single delivery period into several advertisement delivery feature time periods according to the total usage frequency of all software; Determining the advertisement delivery type of the current feature time period according to the usage type of the software with the highest usage frequency in each feature time period; The time length of each advertisement delivery feature time period is different. 3.The user portrait-based advertisement distribution method according to claim 2, wherein, Dividing the advertisement delivery feature time periods into three types according to the total software usage frequency, including a non-active delivery time period, an active delivery time period, and a general delivery time period; If the total usage frequency of all software is less than the non-active frequency threshold, the current advertisement delivery feature time period is a non-active delivery time period; If the total usage frequency of all software is greater than or equal to the active frequency threshold, the current advertisement delivery feature time period is an active delivery time period; If the total usage frequency of the software is greater than or equal to the inactive frequency threshold and less than the active frequency threshold, the current feature time period is a general delivery time period. 4.The user portrait-based advertisement distribution method according to claim 3, wherein, According to the advertisement feedback attitude of the target user in each feature time period, the advertisement effective tendency category of the current user in each delivery feature time period is determined, including: determining the advertisement feedback attitude of the target user each time, the advertisement feedback attitude including positive feedback and negative feedback; determining the advertisement effective tendency category of the current time period according to the proportion of the number of negative feedbacks; If the number of negative feedbacks is greater than the preset negative feedback threshold, the advertisement effective tendency category of the current time period is an advertisement delivery invalid category; If the number of negative feedbacks is less than or equal to the preset negative feedback threshold, the advertisement effective tendency category of the current time period is an advertisement delivery effective category. 5.The user portrait-based advertisement distribution method according to claim 4, wherein, According to the advertisement effective tendency category of the current feature time period, the basic advertisement delivery mode is determined, including: If the advertisement effective tendency category is an advertisement delivery effective category, the basic advertisement delivery mode is to deliver according to the advertisement delivery type of the feature time period; If the advertisement effective tendency category is an advertisement delivery invalid category, the basic advertisement delivery mode is to deliver the advertisement reversely. 6.The user portrait-based advertisement distribution method according to claim 5, wherein, The reverse advertisement delivery includes: In response to the user in the corresponding active delivery time period, an advertisement different from the user information record is delivered, and its advertisement feedback attitude is recorded, and the delivery is replaced in the direction from large to small according to the difference.

Images