Personalized advertisement pushing method based on user interest learning

Abstract

The invention relates to the technical field of advertisement pushing, and discloses a personalized advertisement pushing method based on user interest learning, comprising the following steps: step 1, acquiring basic information of all users and forming a user data set; step 2, acquiring behavior data of all users and forming a behavior data set; step 3, obtaining use scores of all users and forming a questionnaire data set; 4, according to the user data set and the behavior data set, analyzing and generating a daily living data set and daily average use duration, and determining the number of active people of each user group and the daily average use duration of each independent user; and step 5, according to the behavior data set and the questionnaire data set, analyzing and generating an order conversion rate and a satisfaction index, comprehensively analyzing that the user behavior pushing is more accurate, judging the satisfaction degree of each independent user for personalized advertisement pushing, and if the satisfaction degree is low, reminding a platform manager to adjust the pushing content of the personalized advertisement so as to improve the user experience. And the marketing strategy is flexibly adjusted, so that the user experience is better.

Description

Technical Field

[0001] The present invention relates to the technical field of advertisement delivery, and in particular to a personalized advertisement delivery method based on user interest learning. Background Art

[0002] In modern business society, advertising is not only a tool for information dissemination, but also a highly specialized art and science. Advertising delivers goods, services or ideas to the audience through various media, aiming to enhance brand awareness, shape brand image, guide consumer behavior, and ultimately promote sales growth. The core function of advertising is to convey information, attract the attention of the target audience through creative expression, stimulate their interest and desire, and ultimately promote purchasing behavior. Advertising can be divided into print advertising, radio advertising, television advertising, online advertising, etc. Each type of advertising has its own unique characteristics and application scenarios, and choosing the right type of advertising is the key to success. The advertising push method based on user interest learning is a strategy that uses algorithms and data analysis technology to accurately deliver advertising based on user interests, behaviors and preferences. This method collects user browsing history, purchase history, social interaction and other data to build user portraits, thereby pushing advertising content that is highly relevant to user interests. Personalized advertising push methods are conducive to advertisers and users to establish long-term relationships. This relationship not only helps to improve user loyalty, but also brings long-term market returns to brands.

[0003] At present, traditional advertising push methods often ignore the differentiation of individual interests and poor user experience, resulting in low click-through rates and conversion rates of advertisements. In addition, platform managers and advertisers find it difficult to measure advertising effectiveness and cannot accurately determine which marketing strategies are effective and which need to be optimized and improved. Summary of the invention

[0004] 1. Technical issues to be solved

[0005] In view of the shortcomings of the prior art, the present invention provides a personalized advertising push method based on user interest learning, which has the advantages of more accurate push through comprehensive analysis of user behavior, flexible adjustment of marketing strategies and better user experience, etc., which solves the problem that traditional advertising push methods lack individual differentiation and are difficult to measure advertising effects.

[0006] (II) Technical solution

[0007] To achieve the above object, the present invention provides the following technical solution: a personalized advertising push method based on user interest learning, comprising the following steps:

[0008] Step 1: Obtain basic information of all users through the network connection to the big data platform, and form a user data set with the acquired data;

[0009] Step 2: Obtain the behavior data of all users through the network connection to the big data platform, and form the obtained data into a behavior data set;

[0010] Step 3: Obtain the usage scores of all users through the network connection to the big data platform, and form the obtained data into a questionnaire data set;

[0011] Step 4: Analyze and generate the daily active data group Rhsj and the average daily usage time Rjsy based on the user data set and behavior data set, and clarify the number of active users in each user group and the average daily usage time of each independent user in the big data platform;

[0012] Step 5: Analyze and generate the order conversion rate Dzhl and satisfaction index Myzs based on the behavior data set and the questionnaire data set, and determine the satisfaction of each independent user with the personalized advertising push. If the satisfaction level is low, remind the platform administrator to adjust the personalized advertising push content.

[0013] Preferably, in step 1, the expression of the user data set is to Corresponding to the basic information of each user in turn, l represents the user's age, x represents the user's gender, and 1 to n represent that there are n users using the big data platform to browse information.

[0014] Preferably, in step 2, the expression of the behavior data set is {X1 s , X2 s 、X3 s ,...,Xn s}, X1 s To Xn s The corresponding behavior data of each user includes browsing history, search history, comment history, favorite history and purchase history. s represents the duration of the user's continuous behavior in a single day. 1 to n represent that there are n users who use the big data platform to browse information.

[0015] Preferably, in step 3, the expression of the questionnaire data set is {W1 t 、W2 t 、W3 t 、...、wm t}, W1 t To Wm t The corresponding usage scores of each user are shown in sequence. t represents the time point of obtaining the usage score. 1 to m represent that there are m users participating in the questionnaire survey on the big data platform.

[0016] Preferably, in step 4, the calculation process of the daily activity data group Rhsj is as follows:

[0017] Extract the basic information of the i-th user in the user data set and mark it as

[0018] like The age of the user is less than 18 years old, and the i-th user is classified as a minor group WQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as the underage male group WQM. The gender of the user is female, and the i-th user is classified as a minor female group WQF;

[0019] like The age of the users is between 18 and 44 years old, and the i-th user is classified as a young group QQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as a young male group QQM. The gender of the user is female, and the i-th user is classified as a young female group QQF;

[0020] like The age of the middle-aged users is between 45 and 59 years old. The i-th user is classified as the middle-aged group ZQ. The user's gender continues to be classified. If The gender of the user is male, and the i-th user is classified as the middle-aged male group ZQM. The gender of the middle user is female, so the i-th user is classified as the middle-aged female group ZQF;

[0021] like The age of the users is greater than or equal to 60 years old, and the i-th user is classified as the elderly group LQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as the elderly male group LQM. The gender of the user is female, and the i-th user is classified as the elderly female group LQF;

[0022]

[0023] In the formula, Rhsj represents the daily active data group, It means that according to the basic information of each user in the user data set, they are classified into corresponding groups in turn. WQM+WQF represents the total number of underage male groups and underage female groups on the big data platform, QQM+QQF represents the total number of young male groups and young female groups on the big data platform, ZQM+ZQF represents the total number of middle-aged male groups and middle-aged female groups on the big data platform, and LQM+LQF represents the total number of elderly male groups and elderly female groups on the big data platform.

[0024] Preferably, in step 4, the calculation process of the average daily usage time Rjsy is as follows:

[0025] Extract the behavior data of the i-th user in the behavior dataset and mark it as Xi s ;

[0026] Set U day as a single monitoring period H, and during monitoring period H, set Xi s The duration of the continuous behavior of user U in a day is marked as {S1, S2, S3, ..., S U};

[0027]

[0028] In the formula, Rjsy represents the average daily usage time, ∑(S1+S2+S3+...+S U ) represents the total duration of the continuous behavior of the i-th user U days, The total duration of the continuous behavior of the ith user U in a day is divided by the number of days in a single monitoring cycle to obtain the average daily usage time of the ith user.

[0029] Preferably, in step 5, the order conversion rate Dzhl calculation process is as follows:

[0030] Extract the behavior data of the kth user in the behavior dataset and mark it as Xk s ;

[0031] Xk s The total number of times you have viewed personalized ads in your browsing history is marked as Xk s The total quantity of the advertised item in the purchase record is marked as

[0032]

[0033] In the formula, Dzhl represents the order conversion rate. The order conversion rate is the percentage of the total number of advertised products purchased by the kth user to the total number of times the personalized advertisements are viewed.

[0034] Preferably, in step 5, the calculation process of the satisfaction index Myzs is as follows:

[0035] Extract the usage rating of the kth user in the questionnaire dataset and mark it as Wk t ;

[0036] If Wk t There is only a single-use rating in

[0037]

[0038] In the formula, Myzs represents the satisfaction index, and M represents the full score of the big data platform questionnaire survey. Wk t The ratio of the single-use score to the full score is the satisfaction index;

[0039] If Wk t If there are two or more usage ratings in the chronological order from earliest to latest, Wk t Each usage score in is marked as {P1, P2, ..., P f}, 1 to f means that the kth user participated in the questionnaire survey on the big data platform f times, and the corresponding usage score is f times;

[0040]

[0041] In the formula, Myzs represents the satisfaction index, It represents the average score obtained by dividing the sum of all usage scores of the kth user by the number of times he participated in the questionnaire survey. Wk t The ratio of the last usage score to the average score is the satisfaction index.

[0042] Preferably, in step 5, when a single user only has a single usage rating, if the satisfaction index Myzs is less than or equal to , it indicates that the user is not satisfied with the personalized advertising push, and reminds the platform administrator to adjust the personalized advertising push content according to the user's search history, comment history and favorite history in the behavior data set. If the satisfaction index Myzs is greater than , it indicates that the user is highly satisfied with personalized advertising push.

[0043] Preferably, in step five, when a single user has two or more usage ratings, if the satisfaction index Myzs is less than or equal to 1, it indicates that the user is not satisfied with the personalized advertising push, and the platform administrator is reminded to adjust the personalized advertising push content according to the user's search history, comment history and collection history in the behavior data set; if the satisfaction index Myzs is greater than 1, it indicates that the user is highly satisfied with the personalized advertising push.

[0044] Compared with the prior art, the present invention provides a personalized advertising push method based on user interest learning, which has the following beneficial effects:

[0045] 1. The present invention obtains the basic information, behavior data and usage scores of all users through a network connection to a big data platform, and classifies them into user data sets, behavior data sets and questionnaire data sets. According to the user data sets and behavior data sets, the daily active data set Rhsj and the average daily usage time Rjsy are analyzed and generated, and the number of active users in each user group and the average daily usage time of each independent user in the big data platform are clarified, which provides advertisers with important indicators for evaluating advertising effects, helps platform managers to allocate traffic resources more reasonably, and ensures that advertising content can more accurately reach users' potential consumption needs. Then, according to the behavior data sets and questionnaire data sets, the order conversion rate Dzhl and the satisfaction index Myzs are analyzed and generated. The higher the order conversion rate Dzhl, the more effective the current marketing strategy is in attracting users to purchase, focusing on the individual differences of each independent user, and comprehensively analyzing user behavior push more accurately.

[0046] 2. The present invention uses the satisfaction index Myzs to judge the satisfaction of each independent user with personalized advertising push. When a single user only has a single usage rating, if the satisfaction index Myzs is less than or equal to , it means that the user is not satisfied with the personalized advertising push. If the satisfaction index Myzs is greater than When , it means that the user is highly satisfied with the personalized advertising push. When a single user has two or more usage ratings, if the satisfaction index Myzs is less than or equal to 1, it means that the user is not satisfied with the personalized advertising push. This reminds the platform managers to adjust the personalized advertising push content according to the user's search history, comment history and collection history in the behavior data set, and flexibly adjust the marketing strategy for better user experience. If the satisfaction index Myzs is greater than 1, it means that the user is highly satisfied with the personalized advertising push. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a step diagram of the method of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Since traditional advertising push methods often ignore the differentiation of individual interests, the user experience is poor, resulting in low click-through rate and conversion rate of advertisements. In addition, it is difficult for platform managers and advertisers to measure the effectiveness of advertisements and accurately determine which marketing strategies are effective and which need to be optimized and improved. Therefore, a personalized advertising push method based on user interest learning is provided. Please refer to Figure 1 , the method comprises the following steps:

[0050] Step 1: Connect to the big data platform through the network to obtain the basic information of all users, and form the obtained data into a user data set. The expression of the user data set is to Corresponding to the basic information of each user, l represents the user's age, x represents the user's gender, and 1 to n represent that there are n users browsing information on the big data platform. Only collecting the user's age and gender for group classification is helpful for advertisers to locate the target audience more accurately.

[0051] Step 2: Connect to the big data platform through the network to obtain the behavior data of all users, and form a behavior data set with the obtained data. The expression of the behavior data set is {X1 s , X2 s 、X3 s ,...,Xn s}, X1 s To Xn s The behavioral data of each user includes browsing history, search history, comment history, favorite history, and purchase history. s represents the duration of the user's continuous behavior in a single day, and 1 to n represent the number of users browsing information on the big data platform. This provides data support for subsequent analysis of user behavior of different ages and genders, enabling a more accurate understanding of consumer needs and optimization of product design and marketing strategies.

[0052] Step 3: Connect to the big data platform through the network to obtain the usage scores of all users, and form the obtained data into a questionnaire data set. The expression of the questionnaire data set is {W1 t 、W2 t 、W3 t , ..., Wm t}, W1 tTo Wm t The corresponding usage scores of each user are shown in sequence. t represents the time point of obtaining the usage scores. 1 to m represent the number of users who participated in the questionnaire survey on the big data platform. This directly reflects the user's impression of the advertising content and improves the user's participation.

[0053] Step 4: Analyze and generate the daily active data set Rhsj and the average daily usage time Rjsy based on the user data set and behavior data set. The calculation process is as follows:

[0054] Extract the basic information of the i-th user in the user data set and mark it as

[0055] like The age of the user is less than 18 years old, and the i-th user is classified as a minor group WQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as the underage male group WQM. The gender of the user is female, and the i-th user is classified as a minor female group WQF;

[0056] like The age of the users is between 18 and 44 years old, and the i-th user is classified as a young group QQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as a young male group QQM. The gender of the user is female, and the i-th user is classified as a young female group QQF;

[0057] like The age of the middle-aged users is between 45 and 59 years old. The i-th user is classified as the middle-aged group ZQ. The user's gender continues to be classified. If The gender of the user is male, and the i-th user is classified as the middle-aged male group ZQM. The gender of the middle user is female, so the i-th user is classified as the middle-aged female group ZQF;

[0058] like The age of the users is greater than or equal to 60 years old, and the i-th user is classified as the elderly group LQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as the elderly male group LQM. The gender of the user is female, and the i-th user is classified as the elderly female group LQF;

[0059]

[0060] In the formula, Rhsj represents the daily active data group, It means that according to the basic information of each user in the user data set, they are classified into corresponding groups in turn. WQM+WQF represents the total number of underage male groups and underage female groups on the big data platform, QQM+QQF represents the total number of young male groups and young female groups on the big data platform, ZQM+ZQF represents the total number of middle-aged male groups and middle-aged female groups on the big data platform, and LQM+LQF represents the total number of elderly male groups and elderly female groups on the big data platform. The daily active data group includes the number of male and female minors, male and female youth groups, male and female middle-aged groups, and male and female elderly groups. It clarifies the number of active users in each user group on the big data platform, provides advertisers with an important indicator for evaluating advertising effectiveness, and helps platform managers allocate traffic resources more reasonably.

[0061] Extract the behavior data of the i-th user in the behavior dataset and mark it as Xi s ;

[0062] Set U day as a single monitoring period H, and during monitoring period H, set Xi s The duration of the continuous behavior of user U in a day is marked as {S1, S2, S3, ..., S U};

[0063]

[0064] In the formula, Rjsy represents the average daily usage time, ∑(S1+S2+S3+...+S U ) represents the total duration of the continuous behavior of the i-th user U days, The total duration of the continuous behavior of the ith user U divided by the number of days in a single monitoring cycle is the average daily usage time of the ith user. According to the calculation process of the average daily usage time Rjsy, the average daily usage time of each independent user can be analyzed and obtained by analogy, ensuring that the advertising content can more accurately reach the user's potential consumption needs;

[0065] Step 5: Analyze and generate the order conversion rate Dzhl and satisfaction index Myzs based on the behavior data set and questionnaire data set. The calculation process is as follows:

[0066] Extract the behavior data of the kth user in the behavior dataset and mark it as Xk s ;

[0067] Xk s The total number of times you have viewed personalized ads in your browsing history is marked as Xk s The total quantity of the advertised item in the purchase record is marked as

[0068]

[0069] In the formula, Dzhl represents the order conversion rate. The ratio of the total number of advertised products purchased by the kth user to the total number of personalized ads viewed is the order conversion rate. The higher the order conversion rate, the more effective the current marketing strategy is in attracting users to make purchases.

[0070] Extract the usage rating of the kth user in the questionnaire dataset and mark it as Wk t ;

[0071] If Wk t There is only a single-use rating in , which means that the kth user is likely to be a short-term active user and has a low trust in the big data platform.

[0072]

[0073] In the formula, Myzs represents the satisfaction index, and M represents the full score of the big data platform questionnaire survey. Wk t The ratio of the single-use score to the full score is the satisfaction index;

[0074] If Wk t There are two or more usage ratings in Wk, which means that the kth user is likely to be a long-term active user and has a high degree of trust in the big data platform. t Each usage score in is marked as {P1, P2, ..., P f}, 1 to f means that the kth user participated in the questionnaire survey on the big data platform f times, and the corresponding usage score is f times;

[0075]

[0076] In the formula, Myzs represents the satisfaction index, It represents the average score obtained by dividing the sum of all usage scores of the kth user by the number of times he participated in the questionnaire survey. Wk t The ratio of the last usage score to the average score is the satisfaction index, which focuses on the individual differences of each independent user and comprehensively analyzes user behavior to make push notifications more accurate.

[0077] The satisfaction index Myzs is used to determine the satisfaction of each independent user with personalized advertising push. When a single user only has a single usage rating, if the satisfaction index Myzs is less than or equal to , it indicates that the user is not satisfied with the personalized advertising push, and reminds the platform administrator to adjust the personalized advertising push content according to the user's search history, comment history and favorite history in the behavior data set. If the satisfaction index Myzs is greater than When , it means that the user is highly satisfied with the personalized advertising push. When a single user has two or more usage ratings, if the satisfaction index Myzs is less than or equal to 1, it means that the user is not satisfied with the personalized advertising push. This reminds the platform managers to adjust the personalized advertising push content according to the user's search history, comment history and collection history in the behavior data set, and flexibly adjust the marketing strategy for better user experience. If the satisfaction index Myzs is greater than 1, it means that the user is highly satisfied with the personalized advertising push.

[0078] Example 1: In this experiment, a 26-year-old female user was selected as the experimental subject, and cosmetics advertisements were pushed to the user through the big data platform. The monitoring period H corresponding to the user was set to 7 days. According to statistics, during the monitoring period H, the duration of the user's continuous behavior was 3.5h, 2.2h, 1.4h, 2h, 4.5h, 6h, 4.3h, respectively. The calculation formula for the user's average daily usage time is as follows:

[0079]

[0080] In the formula, Rjsy represents the average daily usage time. The total duration of the user's continuous behavior within 7 days is divided by the number of days in a single monitoring cycle, which is 7 days. The average daily usage time is 3.6 hours.

[0081] Example 2: In this experiment, a 46-year-old male user was selected as the experimental subject, and a car advertisement was pushed to the user through the big data platform. According to statistics, the user participated in the questionnaire survey of the big data platform 4 times, and the corresponding usage scores were 65 points, 75 points, 80 points and 90 points respectively. The user's satisfaction index Myzs is calculated as follows:

[0082]

[0083] In the formula, Myzs represents the satisfaction index. The sum of all the user's usage scores divided by the number of times the user participated in the questionnaire survey results in an average score of 77.5 points. The ratio of the last usage score of 90 points to the average score of 77.5 points is 1.2, which is the user's satisfaction index. After judgment, 1.2>1, and the user is highly satisfied with the personalized advertising push.

[0084] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A personalized advertising push method based on user interest learning, characterized by: The following steps are involved: Step 1: Obtain basic information of all users through the network connection to the big data platform, and form a user data set with the acquired data; Step 2: Obtain the behavior data of all users through the network connection to the big data platform, and form the obtained data into a behavior data set; Step 3: Obtain the usage scores of all users through the network connection to the big data platform, and form the obtained data into a questionnaire data set; Step 4: Analyze and generate the daily active data group Rhsj and the average daily usage time Rjsy based on the user data set and behavior data set, and clarify the number of active users in each user group and the average daily usage time of each independent user in the big data platform; Step 5: Analyze and generate the order conversion rate Dzhl and satisfaction index Myzs based on the behavior data set and the questionnaire data set, and determine the satisfaction of each independent user with the personalized advertising push. If the satisfaction level is low, remind the platform administrator to adjust the push content of personalized advertising.

2. The personalized advertising push method based on user interest learning according to claim 1, characterized in that: In step 1, the expression of the user data set is: to Corresponding to the basic information of each user in turn, l represents the user's age, x represents the user's gender, and 1 to n represent that there are n users using the big data platform to browse information.

3. The personalized advertising push method based on user interest learning according to claim 2 is characterized in that: In step 2, the expression of the behavior data set is {X1 s , X2 s 、X3 s ,...,Xn s }, X1 s To Xn s The corresponding behavior data of each user includes browsing history, search history, comment history, favorite history and purchase history. s represents the duration of the user's continuous behavior in a single day. 1 to n represent that there are n users who use the big data platform to browse information.

4. The personalized advertising push method based on user interest learning according to claim 3 is characterized by: In step 3, the expression of the questionnaire data set is {W1 t 、W2 t 、W3 t , ..., Wm t }, W1 t To Wm t The corresponding usage scores of each user are shown in sequence. t represents the time point of obtaining the usage score. 1 to m represent that there are m users participating in the questionnaire survey on the big data platform.

5. The personalized advertising push method based on user interest learning according to claim 4 is characterized by: In step 4, the calculation process of the daily activity data group Rhsj is as follows: Extract the basic information of the i-th user in the user data set and mark it as like The age of the user is less than 18 years old, and the i-th user is classified as a minor group WQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as the underage male group WQM. The gender of the user is female, and the i-th user is classified as a minor female group WQF; like The age of the users is between 18 and 44 years old, and the i-th user is classified as a young group QQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as a young male group QQM. The gender of the user is female, and the i-th user is classified as a young female group QQF; like The age of the middle-aged users is between 45 and 59 years old. The i-th user is classified as the middle-aged group ZQ. The user's gender continues to be classified. If The gender of the user is male, and the i-th user is classified as the middle-aged male group ZQM. The gender of the middle user is female, so the i-th user is classified as the middle-aged female group ZQF; like The age of the users is greater than or equal to 60 years old, and the i-th user is classified as the elderly group LQ. The user's gender continues to be classified. If The user gender is male, and the i-th user is classified as the elderly male group LQM. The gender of the user is female, and the i-th user is classified as the elderly female group LQF; In the formula, Rhsj represents the daily active data group, It means that according to the basic information of each user in the user data set, they are classified into corresponding groups in turn. WQM+WQF represents the total number of underage male groups and underage female groups on the big data platform, QQM+QQF represents the total number of young male groups and young female groups on the big data platform, ZQM+ZQF represents the total number of middle-aged male groups and middle-aged female groups on the big data platform, and LQM+LQF represents the total number of elderly male groups and elderly female groups on the big data platform.

6. The personalized advertising push method based on user interest learning according to claim 5, characterized in that: In step 4, the calculation process of the average daily usage time Rjsy is as follows: Extract the behavior data of the i-th user in the behavior dataset and mark it as Xi s ; Set U day as a single monitoring period H, and during the monitoring period H, set Xi s The duration of the continuous behavior of user U in a day is marked as {S1, S2, S3, ..., S U }; In the formula, Rjsy represents the average daily usage time, ∑(S1+S2+S3+...+S U ) represents the total duration of the continuous behavior of the i-th user U days, The total duration of the continuous behavior of the ith user U in a day is divided by the number of days in a single monitoring cycle to obtain the average daily usage time of the ith user.

7. The personalized advertising push method based on user interest learning according to claim 6 is characterized by: In step 5, the calculation process of the order conversion rate Dzhl is as follows: Extract the behavior data of the kth user in the behavior dataset and mark it as Xk s ; Xk s The total number of times you have viewed personalized ads in your browsing history is marked as Xk s The total quantity of the advertised item in the purchase record is marked as In the formula, Dzhl represents the order conversion rate. The order conversion rate is the percentage of the total number of advertised products purchased by the kth user to the total number of times the personalized advertisements are viewed.

8. The personalized advertising push method based on user interest learning according to claim 7 is characterized by: In step 5, the calculation process of the satisfaction index Myzs is as follows: Extract the usage rating of the kth user in the questionnaire dataset and mark it as Wk t ; If Wk t There is only a single-use rating in In the formula, Myzs represents the satisfaction index, and M represents the full score of the big data platform questionnaire survey. Wk t The ratio of the single-use score to the full score is the satisfaction index; If Wk t If there are two or more usage ratings in the chronological order from earliest to latest, Wk t Each usage score in is marked as {P1, P2, ..., P f }, 1 to f means that the kth user participated in the questionnaire survey on the big data platform f times, and the corresponding usage score is f times; In the formula, Myzs represents the satisfaction index, It represents the average score obtained by dividing the sum of all usage scores of the kth user by the number of times he participated in the questionnaire survey. Wk t The ratio of the last usage score to the average score is the satisfaction index.

9. The personalized advertising push method based on user interest learning according to claim 8, characterized in that: In step 5, when a single user only has a single use rating, if the satisfaction index Myzs is less than or equal to , it indicates that the user is not satisfied with the personalized advertising push, and reminds the platform administrator to adjust the personalized advertising push content according to the user's search history, comment history and favorite history in the behavior data set. If the satisfaction index Myzs is greater than , it indicates that the user is highly satisfied with personalized advertising push.

10. The personalized advertising push method based on user interest learning according to claim 9, characterized in that: In the step 5, when a single user has two or more usage ratings, if the satisfaction index Myzs is less than or equal to 1, it means that the user is not satisfied with the personalized advertising push, and the platform administrator is reminded to adjust the personalized advertising push content according to the user's search history, comment history and collection history in the behavior data set; if the satisfaction index Myzs is greater than 1, it means that the user is highly satisfied with the personalized advertising push.

Images