Intelligent product promotion system based on AI

Abstract

The invention relates to an intelligent product promotion system based on AI, and belongs to the technical field of big data processing and analysis. The system firstly determines the basic weight of each promotion mode according to the proportion of users who purchase promoted products under each promotion mode in historical promotion data, then analyzes the interaction behaviors of the users who purchase the products and the promotion, and determines the reflection weight of each interaction behavior to purchase intentions. And in combination with the reflection weight of each interaction behavior, the specific interaction behavior and the purchase time of the user due to promotion, determining the acceptance speed of the customer for each promotion mode so as to correct the basic weight, and obtaining the optimal relative promotion weight of each promotion mode at present. According to the method, the result dimension and the process dimension are integrated, and the popularization result and the popularization accepting speed are quantitatively considered at the same time, so that the finally obtained relative popularization weight is more accurate in popularization effect characterization of each popularization mode, and the popularization efficiency for the current user platform can be remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of big data processing and analysis technology, and in particular to an AI-based intelligent product promotion system. Background Technology

[0002] Traditional product promotion models suffer from limitations such as low precision and difficulty in evaluating promotion effects. Therefore, with the development of the Internet and the improvement of digitalization, the use of AI technology to analyze and mine massive amounts of user data on user platforms to provide enterprises with feedback information from target customer groups on promotions, enabling enterprises to formulate appropriate promotion strategies, is gradually becoming the mainstream approach.

[0003] While some existing AI-based product promotion systems can achieve a certain degree of personalized recommendations, their recommendation strategies remain relatively simplistic. They typically adjust promotion strategies based on results, lacking a multi-dimensional measurement of customer responses to promotions. For example, they lack a deep understanding and exploration of how well customers accept promotions. Furthermore, user behavior and interest in different promotional methods vary across different platforms. This leads to insufficient consideration of customer feedback on promotions, resulting in ineffective targeting and ultimately low efficiency.

[0004] In other words, current AI-based product promotion systems suffer from low promotion efficiency due to insufficient targeting of the platform. Summary of the Invention

[0005] In view of this, the present invention provides an AI-based intelligent product promotion system to solve the technical problem of low promotion efficiency caused by insufficient targeting of promotion on different platforms.

[0006] The present invention provides an AI-based intelligent product promotion system, comprising:

[0007] The basic weight determination module is used to calculate the percentage of users who purchase the promoted product among those who have been in contact with any promotion method on the current user platform, and use this percentage as the basic weight of that promotion method on the current user platform.

[0008] The weight determination module is used to determine the weight of each promotion response behavior in relation to purchase intention among users who purchase the promoted product after contacting the current promotion method on the current user platform. It determines the purchase intention value of each user under the current promotion response behavior based on the relationship between the value of the current promotion response behavior and purchase intention. It determines the acceptance level of the current promotion method for any user who purchases the promoted product after contacting the current promotion method on the current user platform based on the weight of any promotion response behavior in relation to purchase intention, the purchase intention value of any user under any promotion response behavior, and the purchase time after promotion for any user under the current promotion method. Finally, it adjusts the base weight of the current promotion method on the current user platform based on each user's acceptance level of the current promotion method to obtain the adjusted weight. The purchase time after promotion refers to the time it takes for a user to purchase the promoted product after contacting a certain promotion method, and the promotion response behavior refers to the interaction behavior between the user and the promotion method.

[0009] The promotion implementation module is used to determine the relative promotion weight of a promotion method on the current user platform based on the proportion of users who have been exposed to any promotion method among all users in the current user platform and the adjusted weight of that promotion method, and then complete the subsequent promotion.

[0010] Furthermore, determining the weight of each promotional response behavior in relation to purchase intention includes:

[0011] For users on the current user platform who purchase the promoted product after being exposed to the current promotion method, the value of any promotion response behavior and the purchase time after promotion are fitted with a straight line, and the absolute value of the slope of the fitted line is calculated. The ratio of the absolute value of the slope corresponding to any promotion response behavior under the current promotion method in the current user platform to the sum of the absolute values ​​of the slopes corresponding to all promotion response behaviors is used as the reflection weight of any promotion response behavior under the current promotion method in the current user platform on the purchase intention.

[0012] Furthermore, determining the purchase intention value of each user under the current promotional response behavior includes:

[0013] For promotional response behaviors whose values ​​are directly proportional to purchase intentions, the reciprocal of the sum of the current user's value for any of these promotional response behaviors and a constant 1 is taken as the current user's purchase intention value under that promotional response behavior.

[0014] For promotional response behaviors whose values ​​are inversely proportional to purchase intentions, the reciprocal of the sum of the current user's value for any of these promotional response behaviors and a constant 1 is used as the difference term. The difference obtained by subtracting the difference term from the constant 1 is used as the current user's purchase intention value under this type of promotional response behavior.

[0015] Furthermore, determining the acceptance level of any user on the current user platform who purchases the promoted product after being exposed to the current promotion method includes:

[0016] ,

[0017] in, This represents the acceptance level of the q-th user on the current user platform who purchases the promoted product after being exposed to the v-th promotion method. m represents the number of promotion response behaviors generated by all users on the current user platform who purchase the promoted product after being exposed to the v-th promotion method. This represents the weight of the r-th promotion response behavior on the purchase intention of all users on the current user platform who have been exposed to the v-th promotion method and subsequently purchased the promoted product. Let represent the purchase intention value of the qth user under the rth type of promotion response behavior, T represent the maximum post-promotion purchase time among all users who purchase the promoted product after being exposed to the vth type of promotion, and represent the post-promotion purchase time of the qth user among all users who purchase the promoted product after being exposed to the vth type of promotion.

[0018] Furthermore, the revised weight obtained by modifying the basic weight of the current promotion method on the current user platform includes:

[0019] The corrected weight is obtained by multiplying the average acceptance level of the current promotion method by the base weight of the current promotion method on the current user platform after each user has been exposed to the current promotion method and then purchasing the promoted product.

[0020] Furthermore, determining the relative promotion weight of this promotion method on the current user platform includes:

[0021] The product of the percentage of users in the current user platform who have been exposed to any promotion method and the adjusted weight of that promotion method is used as the promotion effectiveness value of that promotion method in the current user platform.

[0022] The ratio of the effectiveness of any one promotion method on the current user platform to the sum of the effectiveness of all promotion methods on the current user platform is used as the relative promotion weight of any one promotion method on the current user platform.

[0023] Furthermore, it also includes:

[0024] The promotion strategy adjustment launch module is used to determine the original demand for the promoted product on the current user platform based on the frequency of occurrence of various related keywords of the promoted product on the current user platform, the confidence level of various related keywords, and the percentage of users who have actively searched for the promoted product on the current user platform.

[0025] The potential interest of the current user platform in the promoted product is determined by the percentage of users reached by each promotion method in the current user platform among all users.

[0026] The success rate of promoting the product on the current user platform is determined based on the original demand and potential interest. When the success rate is lower than the preset success rate threshold, the basic weight determination module, the corrected weight determination module, and the promotion implementation module are executed.

[0027] Furthermore, determining the initial demand for the promoted product on the current user platform includes:

[0028] The product of the confidence level and the corresponding frequency of occurrence of each relevant keyword under the current user platform is used as the weighted frequency of that relevant keyword, and the average of the weighted frequencies of all relevant keywords is used as the first original demand degree.

[0029] Determine the number of user types in the current user platform. Use the ratio of the number of users in each type who have actively searched for the promoted product to the total number of users in that type as the active search degree of that type of user. Use the sum of the active search degrees of all types of users as the second original demand degree.

[0030] The normalized value of the product of the first and second original demand is taken as the original demand for the promoted product on the current user platform.

[0031] Furthermore, determining the current user platform's potential interest in the promoted product includes:

[0032] The percentage of users on the current user platform who have been exposed to any promotion method is calculated as the current promotion degree of that promotion method. The normalized value of the sum of the current promotion degrees of all promotion methods is used as the potential interest degree of the current user platform in the promoted product.

[0033] Furthermore, determining the success rate of promoting the product on the current user platform based on the original demand and potential interest includes:

[0034] The average of the original demand and the potential interest is used as the success rate of the promoted product on the current user platform.

[0035] The advantages of this invention compared to the prior art are:

[0036] This invention first determines the basic weight of each promotion method based on the percentage of users who successfully purchased the promoted product under each promotion method in historical promotion data. Then, it analyzes the interaction behavior between users who purchased the promoted product after interacting with the promotion method on the current user platform, determining the weight of each promotion response behavior in reflecting purchase intention. Subsequently, it combines the weight of each promotion response behavior, the specific circumstances of the promotion response behavior, and the time it takes for users to purchase the product after receiving the promotion to determine the acceptance speed of each customer for each promotion method, thereby revising the basic weights and obtaining the optimal relative promotion weight for each promotion method on the current user platform at the current moment. This invention integrates the result dimension and the process dimension, simultaneously quantifying and considering the promotion results and the speed of promotion acceptance, making the final relative promotion weights more accurate in representing the promotion effect of each promotion method, thus significantly improving the promotion efficiency for the current user platform. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is a structural block diagram of an AI-based intelligent product promotion system provided in Embodiment 1 of the present invention. Detailed Implementation

[0039] To further illustrate the technical solution of the present invention, specific embodiments are described below.

[0040] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a particular feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. Furthermore, a particular feature, structure, or characteristic in one or more embodiments may be combined in any suitable form, and the terms "comprising," "including," "having," and variations thereof mean "including, but not limited to," unless otherwise specifically emphasized.

[0041] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0042] System Implementation Example:

[0043] See Figure 1 This is a structural block diagram of an AI-based intelligent product promotion system provided in Embodiment 1 of the present invention, including a basic weight determination module 11, a corrected weight determination module 12, and a promotion implementation module 13. The functions of each module are described in detail below:

[0044] The basic weight determination module 11 is used to calculate the percentage of users who purchase the promoted product among users who have been in contact with any promotion method in the current user platform, and use this percentage as the basic weight of that promotion method in the current user platform.

[0045] While existing AI-based product promotion systems, which adjust promotion strategies based on results, cannot maximize promotion efficiency, it is easy to understand that promotion results do reflect the effectiveness of different promotion methods on the user platform to a considerable extent. Therefore, in this embodiment, this step determines the basic weight of each promotion method on the current user platform based on the promotion results generated by each method.

[0046] Specifically, firstly, from all user platforms to be promoted, select any one as the current user platform. This platform can include social and entertainment platforms such as Weibo and Xiaohongshu. Then, during the historical promotion period (i.e., the time period during which the promoted product has been promoted), count the number of users on the current platform who have encountered each promotion method. Then, count the percentage of users who encountered each promotion method and ultimately purchased the target product. The percentage of users who purchased the target product from those who encountered the current promotion method reflects the promotional result of that method on the current platform and is recorded as the base weight. That is, the percentage of users promoted by each promotion method who ultimately purchased the target product is the base weight of that promotion method on the current platform. Thus, the base weight of each promotion method on the current platform can be obtained, and the base weight of the vth promotion method on the current platform is denoted as... .

[0047] The promotion methods are not limited to a few specific types and can include image promotion, link promotion, and video promotion, among others. For example, on the current user platform, if 100 users click on a link promotion and 20 of them choose to purchase the promoted product, 100 users click on an image promotion and 30 of them purchase the promoted product, and 100 users watch a video promotion and 50 of them purchase the promoted product, then the corresponding base weights for these three promotion methods on the current user platform would be 0.2, 0.3, and 0.5, respectively.

[0048] Among them, the behavior records of each user in the current user platform can be automatically captured with the help of AI technology, and the behavior records can be automatically classified according to the promotion method.

[0049] The corrected weight determination module 12 is used to determine the weight of each promotion response behavior in relation to purchase intention among users who purchase the promoted product after contacting the current promotion method on the current user platform, based on the relationship between the values ​​of various promotion response behaviors and the purchase time after promotion. It also determines the purchase intention value of each user under the current promotion response behavior based on the relationship between the values ​​of the current promotion response behavior and purchase intention. Furthermore, it determines the acceptance level of any user who purchases the promoted product after contacting the current promotion method on the current user platform based on the weight of any promotion response behavior in relation to purchase intention, the purchase intention value of any user under any promotion response behavior, and the purchase time after promotion for any user under the current promotion method. Finally, it corrects the basic weight of the current promotion method on the current user platform based on the acceptance level of each user, obtaining the corrected weight. The purchase time after promotion refers to the time it takes for a user to purchase the promoted product after contacting a certain promotion method, and the promotion response behavior refers to the interaction behavior between the user and the promotion method.

[0050] The basic weights of each promotion method obtained in the previous step represent the final promotion result after each promotion method is implemented on the current user platform. However, they do not represent the promotion efficiency of each method. That is, although both promotion methods will eventually lead to the same percentage of promoted users purchasing the target product on the same user platform, there may be cases where users take less time to purchase the target product under one promotion method and more time under the other. Therefore, in order to more clearly define the promotion efficiency of each promotion method on the user platform, so that the promotion system can more rationally allocate the relative placement ratio of different promotion methods on the same user platform and thus improve promotion efficiency, this module analyzes the promotion response behavior of users who were promoted according to different promotion methods in historical periods and eventually purchased the target product, in order to quantify the relative speed at which users accept different promotion methods on the same user platform.

[0051] First, among users who purchase the promoted product after being exposed to the current promotion method on the current user platform, the relationship between the values ​​of various promotion response behaviors and the time taken to purchase after the promotion is used to determine the weight of each promotion response behavior in relation to purchase intention.

[0052] Specifically, different users' reactions to the same promotional method, i.e., their interactive behaviors with the promotional method, will vary. For example, different users may click on the purchase link a different number of times when faced with the same promotional method, such as video promotions, for different users, and the number of purchases or repeat purchases may differ. Therefore, promotional reaction behaviors include, but are not limited to: clicking on the purchase link, browsing the promotional content, adding the promoted product to the shopping cart, etc.

[0053] Therefore, among users who purchase the promoted product after being exposed to the current promotion method on the current user platform, each user who purchases the promoted product, i.e., the target product, will have one or more corresponding promotion response behaviors. Correspondingly, the purchase time after promotion can also be determined. The purchase time after promotion specifically refers to the duration after a user is exposed to a certain promotion method and purchases the promoted product. Each promotion response behavior will have its corresponding value. For example, the value of the promotion response behavior of clicking the purchase link is the number of times the purchase link is clicked, the value of the promotion response behavior of browsing promotional content is the browsing time, and the value of the promotion response behavior of adding the promoted product to the shopping cart is the number of times the promoted product is added to the shopping cart.

[0054] Therefore, by analyzing the data of users who purchased the promoted product after interacting with the current promotion method on the current user platform, the weight of each promotion response behavior in relation to purchase intention can be determined.

[0055] For users on the current user platform who purchase the promoted product after being exposed to the current promotion method, the value of any promotion response behavior and the purchase time after promotion are fitted with a straight line, and the absolute value of the slope of the fitted line is calculated. The ratio of the absolute value of the slope corresponding to any promotion response behavior under the current promotion method in the current user platform to the sum of the absolute values ​​of the slopes corresponding to all promotion response behaviors is used as the reflection weight of any promotion response behavior under the current promotion method in the current user platform on the purchase intention.

[0056] The reason why a linear fit can be achieved between the value of any promotional response behavior and the time to purchase after the promotion is that different users will produce the same promotional response behavior, but the values ​​of the same promotional response behavior will not be the same, and the time to purchase after the promotion will also be different for different users. Since the greater the degree to which a certain promotional response behavior reflects purchase intention, the more obvious the change in the time to purchase after the promotion will be with different values ​​of that promotional response behavior, that is, the larger the absolute value of the slope of the fitted line. Therefore, the weight of different promotional response behaviors in reflecting purchase intention can be determined by the absolute value of the slope of the fitted line.

[0057] The formulaic representation of the weight of any promotion response behavior under the current promotion method on the purchase intention in the current user platform is as follows:

[0058]

[0059] in, This represents the weight of the r-th promotion response behavior on the purchase intention of all users on the current user platform who have been exposed to the v-th promotion method and subsequently purchased the promoted product. Let represent the slope of the fitted line corresponding to the r-th promotion response behavior generated by all users on the current user platform who purchase the promoted product after being exposed to the v-th promotion method. Let m represent the number of types of promotion response behaviors generated by all users on the current user platform who purchase the promoted product after being exposed to the v-th promotion method. This indicates taking the absolute value.

[0060] After determining the weight of each promotional response behavior of a user on the current platform and the current promotional method in relation to their purchase intention, it is also necessary to combine the value of each promotional response behavior with the purchase intention to provide a basis for quantifying the subsequent user acceptance efficiency of each promotional method. Therefore, it is also necessary to determine the purchase intention value of each user under the current promotional response behavior based on the relationship between the value of the current promotional response behavior and purchase intention:

[0061] For promotional response behaviors whose values ​​are directly proportional to purchase intentions, the reciprocal of the sum of the current user's value for any of these promotional response behaviors and a constant 1 is taken as the current user's purchase intention value under that promotional response behavior.

[0062] For promotional response behaviors whose values ​​are inversely proportional to purchase intentions, the reciprocal of the sum of the current user's value for any of these promotional response behaviors and a constant 1 is used as the difference term. The difference obtained by subtracting the difference term from the constant 1 is used as the current user's purchase intention value under this type of promotional response behavior.

[0063] Among them, the value of promotional response behavior that is directly proportional to purchase intention is exemplified by clicking on a purchase link. The more times a user clicks on a purchase link, the stronger their purchase intention is. The corresponding calculation of each user's purchase intention value under this type of promotional response behavior is as follows:

[0064]

[0065] in, This represents the purchase intention value of the q-th user under the r-th type of promotional response behavior, where the value is directly proportional to the purchase intention. This represents the value of the promotion response behavior for the qth user under the r-th type of promotion response behavior.

[0066] Furthermore, an example of a promotional response behavior whose value is inversely proportional to purchase intention is browsing promotional content. The longer a user spends browsing promotional content, the more hesitant they are before purchasing, and the lower their purchase intention. The corresponding calculation of each user's purchase intention value under this type of promotional response behavior is as follows:

[0067]

[0068] in, This represents the purchase intention value of the q-th user under the r-th type of promotional response behavior, where the value is inversely proportional to the purchase intention. This represents the value of the promotion response behavior for the qth user under the r-th type of promotion response behavior.

[0069] Therefore, regardless of whether the value of the promotion response behavior is directly proportional to or inversely proportional to the purchase intention, the purchase intention value of each user under the current promotion response behavior can be obtained and denoted as . Let represent the purchase intention value of the q-th user under the r-th promotion response behavior.

[0070] Subsequently, the acceptance level of any user on the current platform who purchases the promoted product after being exposed to the current promotion method can be determined by the weight of any promotion response behavior on purchase intention, the purchase intention value of any user under any promotion response behavior, and the purchase time of any user under the current promotion method.

[0071]

[0072] in, This represents the acceptance level of the q-th user on the current user platform who purchases the promoted product after being exposed to the v-th promotion method. m represents the number of promotion response behaviors generated by all users on the current user platform who purchase the promoted product after being exposed to the v-th promotion method. This represents the weight of the r-th promotion response behavior on the purchase intention of all users on the current user platform who have been exposed to the v-th promotion method and subsequently purchased the promoted product. Let represent the purchase intention value of the qth user under the rth promotion response behavior, T represent the maximum post-promotion purchase time among all users who purchase the promoted product after being exposed to the vth promotion method, and represent the post-promotion purchase time of the qth user among all users who purchase the promoted product after being exposed to the vth promotion method.

[0073] After obtaining the acceptance level of each user on the current user platform for the current promotion method, the average acceptance level of each user on the current user platform for the current promotion method can be determined as the acceptance efficiency of the current promotion method by users on the current user platform. This allows for the correction of the basic weight of the current promotion method on the current user platform, resulting in the corrected weight.

[0074] The corrected weight is obtained by multiplying the average acceptance level of the current promotion method by the base weight of the current promotion method on the current user platform.

[0075] The formula for obtaining the corrected weights is as follows:

[0076]

[0077] in, Indicates the first The adjusted weight of each promotion method on the current user platform. Indicates the first The basic weight of this promotion method on the current user platform. This represents the average acceptance level of the vth promotion method among all users on the current user platform.

[0078] Therefore, the weights are adjusted. This means that it simultaneously quantifies the promotion effect of the vth promotion method in the current user platform in terms of both the promotion result and the speed at which the promotion is accepted, thus enabling a more accurate representation of the promotion effect of each promotion method in a single user platform based on the basic weights.

[0079] The promotion implementation module 13 is used to determine the relative promotion weight of a promotion method on the current user platform based on the proportion of users who have been exposed to any promotion method in the current user platform and the adjusted weight of that promotion method, and to complete the subsequent promotion.

[0080] Through the implementation of the above modules, the corrected weights of each promotion method in the current user platform have been obtained. However, the final promotion effect is also related to the proportion of users that each promotion method can successfully reach. Therefore, in this module, it is necessary to further combine the corrected weights with the proportion of users that each promotion method can reach in the current user platform to the total number of users to determine the relative promotion weight of each promotion method in the current user platform.

[0081] The product of the percentage of users in the current user platform who have been exposed to any promotion method and the adjusted weight of that promotion method is used as the promotion effectiveness value of that promotion method in the current user platform.

[0082] The ratio of the effectiveness of any one promotion method on the current user platform to the sum of the effectiveness of all promotion methods on the current user platform is used as the relative promotion weight of any one promotion method on the current user platform.

[0083] The formulaic representation of relative promotion weight is as follows:

[0084]

[0085] in, This indicates the relative promotion weight of the v-th promotion method on the current user's platform. Indicates the first The adjusted weight of each promotion method on the current user platform. This represents the percentage of users reached by the vth promotion method on the current user platform out of all users on the current user platform. In other words, it is the ratio of the number of users reached by the vth promotion method on the current user platform to the total number of users on the current user platform. N represents the number of promotion methods.

[0086] Therefore, the relative promotion weight of each promotion method in the current user platform can be determined. This relative promotion weight is based on the analysis of historical promotion feedback data obtained in real time by AI technology, which shows the relative effectiveness of each promotion method in the current user platform. Therefore, it is the most suitable promotion strategy for the current moment. The promotion system can then complete the subsequent promotion of the promoted product in the current user platform according to the relative promotion weight of each promotion method at the current moment.

[0087] It is easy to understand that user behavior and interest in products on the platform change in real time. Therefore, in this embodiment, the acquisition of relative promotion weight needs to be repeated after a set update interval to ensure that the most suitable and efficient promotion strategy is always applied to the promoted product. The value of the set update interval can be set according to the required accuracy of promotion and the amount of computing resources available; this embodiment does not specify a particular value.

[0088] This invention first determines the basic weight of each promotion method based on the percentage of users who successfully purchased the promoted product under each promotion method during historical promotion periods. Then, it analyzes the interaction behavior between users who purchased the promoted product after interacting with the promotion method on the current user platform, determining the weight of each promotion response behavior in reflecting purchase intention. Subsequently, it combines the weight of each promotion response behavior, the specific circumstances of the promotion response behavior, and the time it takes for users to purchase the product after receiving the promotion to determine the acceptance speed of each customer for each promotion method, thereby revising the basic weights and obtaining the optimal relative promotion weight for each promotion method on the current user platform at the current moment. This invention integrates the result dimension and the process dimension, simultaneously quantifying and considering the promotion results and the speed of promotion acceptance, making the final relative promotion weights more accurate in representing the promotion effect of each promotion method, thus significantly improving the promotion efficiency for the current user platform.

[0089] Embodiment 2 of the present invention also provides an AI-based intelligent product promotion system, which, in addition to the basic weight determination module 11, the corrected weight determination module 12, and the promotion implementation module 13 included in Embodiment 1, further includes a promotion strategy adjustment and initiation module 10.

[0090] The reason for setting up an additional promotion strategy adjustment initiation module 10 in this embodiment is that adjusting the promotion strategy requires a lot of resources. Also, since the user situation is different on different platforms, the users on the current user platform generally have a high demand for the promoted product and can achieve a good promotion effect without vigorous promotion. Or the promotion strategy currently implemented may be an adjusted and optimized promotion strategy based on the relative promotion weight obtained in the above embodiment. Therefore, in order to avoid unnecessary frequent adjustments to the promotion strategy, this embodiment adds the promotion strategy adjustment initiation module 10 to first determine whether the promotion strategy needs to be adjusted. Only when it is determined that adjustment is needed will the operation contents of the basic weight determination module 11, the corrected weight determination module 12, and the promotion implementation module 13 be executed.

[0091] The promotion strategy adjustment launch module 10 is used to determine the original demand for the promoted product on the current user platform based on the number of times various related keywords of the promoted product appear on the current user platform, the confidence of various related keywords, and the proportion of users who have actively searched for the promoted product on the current user platform.

[0092] The potential interest of the current user platform in the promoted product is determined by the percentage of users reached by each promotion method in the current user platform among all users.

[0093] The success rate of promoting the product on the current user platform is determined based on the original demand and potential interest. When the success rate is lower than the preset success rate threshold, the basic weight determination module, the corrected weight determination module, and the promotion implementation module are executed.

[0094] The preset success rate threshold can be set according to the requirements for the success rate of promotion, and this embodiment does not specifically limit it. Furthermore, the confidence level of all keywords associated with the promoted product in the current user platform is obtained using the Apriori algorithm. The Apriori algorithm is existing technology and will not be described in detail here.

[0095] Furthermore, preferably, the success rate of the promotion is:

[0096]

[0097] in, This indicates the success rate of promoting the product on the current user's platform. This indicates the number of times the i-th relevant keyword related to the promoted product appears in the current user platform. This represents the confidence level of the i-th relevant keyword in the current user platform, where n represents the number of relevant keyword types for the promoted product in the current user platform. This indicates the proportion of users in category f on the current user platform. Categories such as female users, male users, young users, and middle-aged users can be identified and categorized using AI based on platform characteristics. For example, WeChat can categorize users by age, and Xiaohongshu can categorize them by interests. This represents the number of users in user category f who have actively searched for the promoted product. Indicates the first level of primary demand. Indicates the second primary demand level. This indicates the current level of initial demand for the promoted product on the user platform. This represents the percentage of users reached by the v-th promotion method on the current user platform, which is also the current promotion effectiveness of the v-th promotion method. N represents the number of promotion methods. This indicates the current user's level of interest in the promoted product on the platform. Normalization refers to methods such as linear normalization, norm normalization, etc.

[0098] This embodiment can first determine the current promotion effect of the promoted product on the current user platform based on the aforementioned embodiment one, and update the promotion strategy when it is determined that the promotion effect is poor, so as to avoid unnecessary frequent changes in promotion strategy, reduce the waste of promotion resources, and further improve promotion efficiency.

[0099] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. An AI-based intelligent product promotion system, characterized in that, include: The basic weight determination module is used to calculate the percentage of users who purchase the promoted product among those who have been in contact with any promotion method on the current user platform, and use this percentage as the basic weight of that promotion method on the current user platform. The weight determination module is used to determine the weight of each promotion response behavior in relation to purchase intention among users who purchase the promoted product after contacting the current promotion method on the current user platform. It determines the purchase intention value of each user under the current promotion response behavior based on the relationship between the value of the current promotion response behavior and purchase intention. It determines the acceptance level of the current promotion method for any user who purchases the promoted product after contacting the current promotion method on the current user platform based on the weight of any promotion response behavior in relation to purchase intention, the purchase intention value of any user under any promotion response behavior, and the purchase time after promotion for any user under the current promotion method. Finally, it adjusts the base weight of the current promotion method on the current user platform based on each user's acceptance level of the current promotion method to obtain the adjusted weight. The purchase time after promotion refers to the time it takes for a user to purchase the promoted product after contacting a certain promotion method, and the promotion response behavior refers to the interaction behavior between the user and the promotion method. The promotion implementation module is used to determine the relative promotion weight of a promotion method on the current user platform based on the proportion of users who have been exposed to any promotion method among all users in the current user platform and the adjusted weight of that promotion method, and then complete the subsequent promotion.

2. The AI-based intelligent product promotion system according to claim 1, characterized in that, The determination of the weight of each promotional response behavior in relation to purchase intention includes: For users on the current user platform who purchase the promoted product after being exposed to the current promotion method, the value of any promotion response behavior and the purchase time after promotion are fitted with a straight line, and the absolute value of the slope of the fitted line is calculated. The ratio of the absolute value of the slope corresponding to any promotion response behavior under the current promotion method in the current user platform to the sum of the absolute values ​​of the slopes corresponding to all promotion response behaviors is used as the reflection weight of any promotion response behavior under the current promotion method in the current user platform on the purchase intention.

3. The AI-based intelligent product promotion system according to claim 1, characterized in that, Determining the purchase intention value of each user under the current promotion response behavior includes: For promotional response behaviors whose values ​​are directly proportional to purchase intentions, the reciprocal of the sum of the current user's value for any of these promotional response behaviors and a constant 1 is taken as the current user's purchase intention value under that promotional response behavior. For promotional response behaviors whose values ​​are inversely proportional to purchase intentions, the reciprocal of the sum of the current user's value for any of these promotional response behaviors and a constant 1 is used as the difference term. The difference obtained by subtracting the difference term from the constant 1 is used as the current user's purchase intention value under this type of promotional response behavior.

4. The AI-based intelligent product promotion system according to any one of claims 1 to 3, characterized in that, The determination of the acceptance level of any user on the current user platform who purchases the promoted product after being exposed to the current promotion method includes: ， in, This represents the acceptance level of the q-th user on the current user platform who purchases the promoted product after being exposed to the v-th promotion method. m represents the number of promotion response behaviors generated by all users on the current user platform who purchase the promoted product after being exposed to the v-th promotion method. This represents the weight of the r-th promotion response behavior on the purchase intention of all users on the current user platform who have been exposed to the v-th promotion method and subsequently purchased the promoted product. Let represent the purchase intention value of the qth user under the rth promotion response behavior, T represent the maximum post-promotion purchase time among all users who purchase the promoted product after being exposed to the vth promotion method, and represent the post-promotion purchase time of the qth user among all users who purchase the promoted product after being exposed to the vth promotion method.

5. The AI-based intelligent product promotion system according to claim 1, characterized in that, The revised weight obtained by correcting the base weight of the current promotion method on the current user platform includes: The corrected weight is obtained by multiplying the average acceptance level of the current promotion method by the base weight of the current promotion method on the current user platform after each user has been exposed to the current promotion method and then purchasing the promoted product.

6. The AI-based intelligent product promotion system according to claim 1, characterized in that, Determining the relative promotion weight of this promotion method on the current user platform includes: The product of the percentage of users in the current user platform who have been exposed to any promotion method and the adjusted weight of that promotion method is used as the promotion effectiveness value of that promotion method in the current user platform. The ratio of the effectiveness of any one promotion method on the current user platform to the sum of the effectiveness of all promotion methods on the current user platform is used as the relative promotion weight of any one promotion method on the current user platform.

7. The AI-based intelligent product promotion system according to claim 1, characterized in that, Also includes: The promotion strategy adjustment launch module is used to determine the original demand for the promoted product on the current user platform based on the frequency of occurrence of various related keywords of the promoted product on the current user platform, the confidence of various related keywords, and the percentage of users on the current user platform who have actively searched for the promoted product. The potential interest of the current user platform in the promoted product is determined by the percentage of users reached by each promotion method in the current user platform among all users. The success rate of promoting the product on the current user platform is determined based on the original demand and potential interest. When the success rate is lower than the preset success rate threshold, the basic weight determination module, the corrected weight determination module, and the promotion implementation module are executed.

8. The AI-based intelligent product promotion system according to claim 7, characterized in that, Determining the initial demand for the promoted product on the current user platform includes: The product of the confidence level and the corresponding frequency of occurrence of each relevant keyword under the current user platform is used as the weighted frequency of that relevant keyword, and the average of the weighted frequencies of all relevant keywords is used as the first original demand degree. Determine the number of user types in the current user platform. Use the ratio of the number of users in each type who have actively searched for the promoted product to the total number of users in that type as the active search degree of that type of user. Use the sum of the active search degrees of all types of users as the second original demand degree. The normalized value of the product of the first and second original demand is taken as the original demand for the promoted product on the current user platform.

9. The AI-based intelligent product promotion system according to claim 7 or 8, characterized in that, Determining the current user platform's potential interest in the promoted product includes: The percentage of users on the current user platform who have been exposed to any promotion method is calculated as the current promotion degree of that promotion method. The normalized value of the sum of the current promotion degrees of all promotion methods is used as the potential interest degree of the current user platform in the promoted product.

10. The AI-based intelligent product promotion system according to claim 9, characterized in that, Determining the success rate of promoting a product on the current user platform based on the original demand and potential interest includes: The average of the original demand and the potential interest is used as the success rate of the promoted product on the current user platform.

Images